GripForge MCP Server

combat.abilities: inspect schema, get an editable Chrono Crab example, validate a pack or export executable Godot, Unity and Unreal source adapters. Shared costs, cooldowns, conditions, targeting, windup/active/recovery and presentation cues. Native adapters require the game combat backend and asset bindings. Returns relative path/content files; no rig/animation/VFX generation or automatic installation. Persist through gripforge_gamekit_configure config.pack or project data. 0 credits.

A hero's abilities ready to play in Character Studio: the kit (role, lane, basic attack, passive + Q/W/E/R with type, range, cooldown, cost, effect, VFX brief, cast animation) from a hero concept or written from one sentence; the basic attack and Q/W/E/R casts animated with NVIDIA Kimodo on the rigged character; and the gripforge.abilities/1 pack (targeting, damage/heal, timings) bound to those clips and stored on the result. Call stage=plan first (free) and show the abilities and price (1 credit per clip, 5 for a kit); then stage=build with the returned kit. Poll gripforge_generation_read with job_id; result.studio_url opens the Abilities tab. Numbers are starting values, not a balance pass; VFX lines are briefs for gripforge_vfx_generate.

Journey — after the roster; gripforge_game_audit lists it among what a finished MOBA has. Write and bind a cast effect for each ability of a Game Kit project that has none: an area gets a ground shockwave, a line a bolt and its trail, an arc a sweeping crescent, a self cast a rising aura; colour from the ability's words (fire, frost, wind, holy…), size from its range, ultimates larger. The project needs the fx.ability_vfx kit (the moba preset installs it) and an abilities collection (gripforge_moba_roster writes it). Effects are procedural sources saved as VFX items and bound to ability_vfx:<ability id>, immediately. replace: true rewrites the existing ones; abilities limits to some ids. A hand-made or generated VFX bound to the same slot takes over. 0 credits.

ADMIN ONLY. Every asset of the platform, all workspaces: filter by kind, owner, tag, pack, style, team rating (rating=none is the review queue) or curation (featured, hidden), sort by rating or size. Each row has its team rating (1–5), featured / hidden flags and its mobile rank (0–100). Pass id for one asset in full. Refused for non-admin keys. Change them with gripforge_admin_assets_update.

ADMIN ONLY. Curate platform assets in bulk (up to 200 ids): rate 1–5 (moves the community rank), add or remove a tag, replace the tags of one asset, feature (first in the rank sort), hide (out of listings and search, still served by id), publish / unpublish, file into a pack. Every change is audited with the admin who made it. Find ids with gripforge_admin_assets. Deleting stays in the /admin/assets screen. Refused for non-admin keys.

Aim-offset poses for a Library character/armed bind (0 credits): per hand, two-bone IK + wrist alignment so the WEAPON AXIS (gf_muzzle locator from loadout, else the seat forward) points at every direction of a yaw×pitch grid. Returns bone-local quaternions per pose + rest, and a Godot blend snippet. Dual wield ready (independent hands).

Retarget the Universal Animation Library pack onto a Mixamo-named Library character/enemy/armed bind. In a Game Kit project gripforge_gamekit_creatures runs it for free on every bound creature, and gripforge_game_audit lists the models still lacking movement or action clips. Base slots + extras (taunt, stinger, knockback, block, jump_start/loop/land, idle_guns, shoot). archetype remaps the base slots: sword (default), claws, heavy, brawler, puppet. Opt-in slots fire_breath (mouth exhalation) and energy_cast (two-hand azure wave) use the Studio VFX character motions. For captured two-hand saber guards/parries use actions with pack=touche and TC_Saber_* clips from GET /api/v1/animate-pack; retains CC BY 4.0 attribution. Explicit actions replace the preset selection. Saves kind=animation. 0 credits.

Plan or apply render.realism across shared PBR materials, immutable textures, surface shaders, lighting, post-processing, environment and skinned character rendering, with a bounded render quality and render.lod configuration suggestion. Defaults to mode=plan (no save); mode=apply writes one work revision atomically, preserving assets, transforms, animation and validated versions. Explicit nodeIds restrict instance assignments; lighting/post-processing are global. materials/characters rows use {nodeId,material:"*"|exactName,override:{roughness,metalness,map,normalMap,surface,...}}; pinned map references, null removes overrides. Characters require character/enemy/boss/fps-arms role. Does not fabricate missing detail, execute paid generation, add game LOD automatically, or approve realism visually. Request a real scene_review afterwards. WEB only; not DLSS.

Read the reusable Meshy/Tripo building and district contract, example, limits and plan → build workflow. Separate immutable Library assets, scene instances and shared SceneDocument. No paid generation.

Queue a durable armor-kit job: one shared concept, isolated Imagine references, Meshy PBR models, bounded Blender interior preparation and paired GLB exports, then actual GripForge wearer review. Types: helmet, chest, belt, pauldron, bracer, glove, thigh, greave, boot, upperarm, gorget, tasset, undersuit. coverage=full adds a durable anatomical decomposition of the concept, rejects missing planned roles before mesh generation and records per-instance coverage fits. A piece count alone never proves body coverage. The optional undersuit is a locally manufactured adaptive textile which retains each wearer’s skin weights. Meshy pieces cost 11 credits per canonical type; the textile has no image/mesh charge. A new concept costs 1 credit; mirrors are included. concept_item must be an owned Library image. source_kit + explicit pieces regenerates only those types into a NEW work kit and pins unchanged pieces. Closing the call/page does not cancel. Poll gripforge_generation_read. Success is a work version, never automatic visual approval. The result includes a Grip Studio link, review scene, concepts, source models and Blender source.

Read an owned kit manifest, semantic slots and immutable piece revisions. Equipment and instances remain independent. Returns the Grip Studio link. To correct selected types, call gripforge_armor_generate with source_kit and pieces.

Fork a failed/cancelled armor job into a new work kit. Preserve completed pieces and the exact shared concept; recreate only unfinished types by default. Optional pieces chooses types explicitly. Use prompt for targeted corrections. Original checkpoints and artifacts remain intact. Costs 11 credits per regenerated canonical type (plus a concept credit only if none survived). Use generation_retry instead for a transient provider or worker error to reuse the exact same steps without another charge.

Copy a published Community asset into the connected workspace. Assets and armor sets are free by default. Only an explicit creator price is charged, using paid credits. Repeated takes are not billed again but may create another copy: reuse an existing workspace id when possible. Returns the private Library item and signed file URL. Download it with gripforge_library_pull, then import into the actual game.

Search actual asset models, materials and animations, independently of Game Kit modules. Returns provenance, import requirements and Library ids. Results are ranked best first by fitness for a GripForge game (score/100, grade A–D, reasons: triangle budget for the target, rigged and clean rig, clips, empty-handed characters, attachable weapons, classified, file size, adoption); `best` is the top pick across sources. Prefer grade A/B; read the reasons before choosing a C/D. Search both sources before manufacturing missing roles. An empty kit search does not mean there are no assets. Search is free; never substitutes proxy geometry for finished models. In a Game Kit project, use it to fill what gripforge_game_audit reports missing (a creature slot gets ranked proposals from gripforge_gamekit_creatures; a character without clips is animated, not left idle).

Attach a prop (sword, shield, gun, staff/scythe…) onto a rigged character. Finds the hand bone across naming schemes, scales the prop to the character, seats the grip, closes the fist, and returns a bind + ready-to-paste Unity / Godot snippets. For melee, provide prop_grip.edgeAxis (signed cutting direction, distinct from the flat normal) and handleSpan (usable handle length). Read bind.orientation and bind.contact; missing or ambiguous measurements remain needs_review. Pass Library ids (character_id / prop_id) or https URLs. attach_id reuses an existing armed bind GLB as the character. prop_id_2 attaches a second held item (off-hand) onto the first result. The bind (+ armed GLB when export_glb) is saved in Library — not a second credit. To write files into an open repo, use the npm MCP package gripforge_library_pull.

Generate SFX and ambience with self-hosted Stable Audio 3 Medium, the default sound model. No downgrade to Small. Call action=quote first (no generation/charge), then action=generate with the returned quote_hash and identical inputs. Cost is disclosed in the quote; generation may spend the selected AI allowance or GripForge credits. Returns a durable job_id immediately; poll gripforge_generation_read. Successful takes survive retries as Library audio with prompt, seed, revision and WAV measurements. Optional role assignment uses expected_revision; a conflicting edit keeps takes in Library without overwriting the project. Listen to results: quality and seamless loops are unreviewed. Max SFX 30s, ambience 60s, 1–4 variants, seed increments per take. Music uses its own pipeline; voices use Drama 3 via audio.speech.

Search locker + Community audio. Prompt matches name, description, tags and use (gunshot, ui, magic, footsteps, announcer, dinosaure, magie, pas…). Each hit has id, description, use[], pack, file_url. Community hits: clone with /api/v1/community/{id}/clone (free unless the creator sets a price in credits) then pull. Free to search.

Inspect project structure and flow, preserve fresh build evidence and report unsupported tests explicitly. 0 credits.

Queue a bounded Blender recipe, never arbitrary Python. Saves GLB and .blend in the workspace as an unvalidated work version with a shared Scene Studio link. Read gripforge_scene_schema for recipe fields. Supports original primitive/custom meshes, PBR colors and keyed transforms. Poll gripforge_generation_read; final visual approval must use GripForge rendering, not Blender screenshots.

Create a playable boss for a game look: stats, phases, attacks bound to anim-pack slots, arena, and Godot / Unity / Unreal snippets. Kit is 0 credits. Reuses a locker enemy when one exists. Pass character_id to bind a Library mesh. Pass generate=true to forge a NEW kind=enemy boss mesh (10 credits, 2–8 min) — only if style_kit / locker is empty. Then gripforge_animate(archetype) + gripforge_hitbox. Name stays text.

Navmesh bots as server-owned player scenes so spawner/synchronizer/net-kit damage apply unchanged. 0 credits.

Not implemented. Returns 501 until artifact upload exists.

Read one game build. Upload/deploy of Linux dedicated artifacts is not implemented yet.

List game_builds (not Godot kit zips at /api/v1/builds).

Not implemented. Returns 501. Mutating / production-sensitive.

Local optional Bevy neural snapshot, not a saved scene edit. Requires an open authorized browser runtime session. Configure off/preview; realtime_experimental is refused. Trained weights and a local rights manifest must be selected in the browser; none are downloaded. render_preview starts an asynchronous request: read the command receipt, then query get_render_capabilities for actual completion, model, effective resolution, provenance, timings and limitations. Cancellation prevents stale publication. No images are uploaded. 0 credits.

Cape rig v2: generates a 3-column bone grid (gf_cape_c*r*) anchored to the neck/upper-spine bone, skins the cape mesh onto it (vertical blend × gaussian columns), and bakes attached_cape.glb (character + skinned cape). The cape then ANIMATES: run the returned Godot script (spring bones, attach to the Skeleton3D) — the engine simulates, the bind carries structure + suggested spring params. Inputs: Library ids (lib_…) or https URLs of GLBs. CONSUMER RECIPE — (1) call with character_id/cape_id; (2) download the baked glb from library.file_url; (3) drop cape-godot.gd next to it and attach to the Skeleton3D node. 1 credit.

Three LOOK pictures (3/4, Imagine low 1K) of a named character, enemy or boss BEFORE any 3D. Pass weapon to show it on the looks; 3D later is T-pose empty hands. Pick one, then gripforge_generate_character with concept_item. 1 credit per image (default 3).

Queue bounded completion using existing kit installers and regression tests. Installs only the provider that fits the game and only kits of its genre, preset or request (never a kit that makes it another game). Preserves exclusions and project revision safety. No paid generation or destructive replacement. Poll status_url; succeeded means the job finished, inspect report.ready. 0 credits.

DEPRECATED alias of gripforge_library_list with engine=compute (each item carries the same `compute` block) and of gripforge_library_get (its `compute` block: family and edit link). List recent GripForge Compute assets of the workspace (manufactures and variants), newest first, each with asset_id, kind, type, category, tags (engine, recipe, budget, near/scene, materials, style, licence…), needs_rig, level (scene/detail), edit_url and file_url. With asset_id: that asset's family (original + colour/material versions) and its edit link. Free.

DEPRECATED alias of gripforge_library_get (same signed library file). Signed-in download link of one Compute GLB (file_url, needs the API key header) and its edit page. Free.

DEPRECATED alias of gripforge_generate_prop with provider=compute (same request through the single generation door POST /api/v1/generate-asset, same answer); prefer that tool. ON DEMAND ONLY: use when the user asks for GripForge Compute; the default 3D generator stays gripforge_generate_prop without provider. Make textured 3D prop/furniture assets with GripForge Compute from images or text descriptions: one asset (name + prompt or image) or several at once (items, up to 5; each item is one asset, from its prompt or its image). Two steps: stage=quote (free) returns each quote.token and the total price; show it to the user, then stage=generate with the tokens (quote_token, or quote_tokens in item order) and confirm=true launches one durable job per asset, each debited once after success with its own idempotency key. Text assets stop after four reference images (billed on their own): show them to the user, then gripforge_compute_validate with the chosen image (the 3D is quoted then, and billed when finished), or regenerate; auto_pick=true (only if the user asked) takes the first image and bills images + 3D once. Weapons are refused before any charge; characters, mobs, creatures and vehicles need their type. Returns one job per asset with studio_url; poll gripforge_generation_read (or this alias family: gripforge_compute_job) for the GLBs and their edit_url.

DEPRECATED alias of gripforge_generation_read (same GET /api/v1/generation-jobs/:id; that tool returns the same studio_url, attente, validation and result.assets edit_url / file_url). Read GripForge Compute jobs (generations or variants): job_id, or job_ids for several at once. Status (queued, running, succeeded, failed), current stage and stage counts. While the router waits, `attente` = {raison: demarrage_machine | aucune_machine | file | machine_occupee, position, depuis, estimation_s} (a machine start takes about 10 min the first time). On success, result.assets lists the GLBs (scene, detail) with asset_id, triangles, edit_url (open it to edit in the studio) and file_url. Free.

GripForge Compute readiness (free, read-only): whether generation is open, which actions the router can run now (image, text, couleur, matiere) and a machine summary (ready / busy / starting). The router rents and stops its own machines; nothing is started by this call.

A text generation of GripForge Compute stops after its four reference images (gripforge_generation_read shows validation.etat en_attente and images[{stage, url}]). Show them to the user, then: image=1..4 to turn that image into 3D (stage=quote returns the 3D price; stage=generate with quote_token and confirm=true launches the 3D job, billed when finished), or regenerate=true (optionally with a new prompt) for four new images, billed again on their own, or abandon=true (only the images were billed).

New version of a Compute asset, on demand: kind=couleur (value #RRGGBB, recolours the main material), kind=matiere (value: a few words, e.g. "green velvet"; the texture is redone) kind=retexture (value: a style for the whole asset, same mesh and UVs, e.g. "painted blue wood, worn") or kind=budget (change the polygon count: value bas | moyen | haut relative to the current one, or an exact « near » triangle count 500–200000; the « scene » GLB gets half; refused as coming soon until the router supports it). stage=quote (free) returns quote.token; after the user accepts, stage=generate with quote_token and confirm=true launches one billed job. Poll gripforge_compute_job; new assets come back with edit_url. Painted-area edits need the studio brush: give the user the edit_url.

Turn a character concept image into a strict T-pose sheet (optional: remove flowing cape/train only — sleeved jacket stays; outfit or sport base). Saves PNG to Library. 1 credit. Hosted MCP takes https file_url.

Local optional Bevy neural snapshot, not a saved scene edit. Requires an open authorized browser runtime session. Configure off/preview; realtime_experimental is refused. Trained weights and a local rights manifest must be selected in the browser; none are downloaded. render_preview starts an asynchronous request: read the command receipt, then query get_render_capabilities for actual completion, model, effective resolution, provenance, timings and limitations. Cancellation prevents stale publication. No images are uploaded. 0 credits.

Queue multiview mesh inspection and vision anatomy detection for an OWNED GLB. Detect actual body, legs, arms, claws, wings or tail without imposing a humanoid/eight-leg template. Returns a persistent job_id. Poll gripforge_generation_read with include_preview=true for the annotated views, editable profile and uncertainty flags. No rig mutation. 0 GripForge credits; configured server vision/Blender required.

Create a NEW private draft character/enemy from a completed creature analysis and optional corrected anatomy. Blender builds the skeleton, surface-smoothed skin weights and procedural starting/diagnostic clips. Grounded leg chains get in-place walk/run with IK. Returns job_id; poll gripforge_generation_read for the Character Studio link, skinned GLB, editable .blend, anatomy and deformation report. Never automatically replaces or validates the source. Every anatomy proposal requires explicit review; use generation_cancel/retry to control durable work. Finger/facial authoring and finished combat choreography are outside v1. 0 GripForge credits.

Read the reusable creature anatomy/rig workflow, coordinate contract, limits and export formats. Detect anatomy before authoring; source GLB, asset instances and draft rig stay separate. Blender is the worker, GripForge is the review renderer.

List the play tabs of a game that a member of the workspace has open right now (heartbeat within 30 s; closed tabs are gone): session id, last seen, viewer role, engine, backend (webgpu/webgl2) and render thread (worker/main). The first is the default target of gripforge_debug_snapshot / gripforge_debug_set. Empty: no tab is open, and the debug tools answer at once that no game tab is open. 0 credits.

Send ops of gripforge.debug/1 to the debug inspector of a game open in a browser by a member of its workspace, then read the snapshot: { op: "mode", wireframe?, colliders?, bounds? } (wireframe, collision boxes, mesh bounds), { op: "tree", root?, depth?, maxNodes? }, { op: "select", ids }, { op: "hide", id, hidden }, { op: "outline", id, on }, { op: "reset" } (show everything, no outline), { op: "extension", id, on } (a debug mode of a kit, e.g. ai.navigation.paths), { op: "open" } / { op: "close" } (the panel on screen). Ids are those of the snapshot tree. Nothing is saved in the project. Same rights as the panel: members of the workspace. 0 credits.

Read the debug inspector of a game open in a browser by a member of its workspace (the play page, Bevy player): frame rate and frame time (p50/p95/p99, worst, slow frames, stages of a frame), GPU and backend (webgpu, webgl2, vulkan, dx12, metal…, software rasteriser or not), draws, meshes, triangles, materials, lights, textures and their memory, entities, memory; the modes (wireframe, colliders, bounds), the selection, hidden and outlined nodes, the debug modes the kits registered and their data; with tree=true the scene tree (scenes, nodes, meshes with vertices / triangles / material, world boxes) within depth and max_nodes. JSON contract gripforge.debug/1, identical on every engine. The inspector is open to everyone by default; a game where the developer turned it off (kit debug.inspector config enabled: false, set with gripforge_gamekit_configure) answers "not open to this viewer". Use it to find why a game is slow or wrong before changing it. Same rights as the panel: members of the workspace. 0 credits.

Journey — once the user's game is understood (any game: no character, enemy, level, ending or known genre is required). Record what the game contains; for a known genre, its definition of done and the common frame (title / pause / end screens, sound, HUD, touch controls, a way to win or lose when the game has an end) are proposals to offer, recorded when the user takes them. Content is data, not code: generic systems in code (kits, parameters, hooks); levels, characters, weapons, items, balance, texts, assets and inputs as data read at run time; clips, bones, grips and sizes from each asset's passport, never written in code. Persist a versioned Game Completion Contract with revision checking. Read get_completion_report for its schema and a proposed contract. Required, optional and excluded capabilities describe the actual intended experience. 0 credits.

The GripForge docs, embedded in this MCP (readable offline, versioned with it), as Markdown. Without topic: the list of topics. With topic: its text, or one section (section = a heading, its number such as "4.2", or words of it); a long topic answers its table of contents first unless a section is given or full=true. Topics: web-export, godot, unity, unreal, desktop, rooms-server, central-server, c-api, lod, shadows, collision, render-lod, render-optimize, core-loop, local-project, game-plan, warnings, physics-async, structs, net, web-multiplayer, scenes, input, combat, character, enemies, npcs, quests, inventory, economy, vehicles, puzzles, sport, ui-webui, ui-screens, world-terrain, minion-waves, game-cinematic, debug-inspector, world-environment, hud, audio-core, i18n-text, save-persistence, settings-graphics, asset-preload, ai-navigation, fx-kits, glossary, kit-api, third-party, player, render-budget, scene-optim, porting, game-schema, module-calls, module-calls-native, godot-host, kit-modules, kit-structures, kit-third-party, abi, modding, game-engine-contract, game-engine, animal-behavior, animal-breeding, animal-taming, asset-cache, asset-meta, automation-logistics, bots-skirmish, build-modular, build-placement, camera-strategy, cards-deck, central-client, character-first-person, character-forms, character-platformer, character-roster, character-top-down, checkpoint-rest, collect-tokens, collection-packs, combat-abilities, combat-attacks, combat-defense, combat-duel, combat-grabs, combat-hit-reactions, combat-lockon, combat-melee, combat-moveset, combat-projectiles, combat-ranged, combat-stamina, combat-threat, combat-turns, combat-weakpoints, companion-control, coop-session, craft-recipes, custom-paint, defense-towers, defense-waves, desktop-launcher, dialogue-voice, display-fullscreen, drop-parachute, encounters-waves, enemy-boss, enemy-horde, enemy-stalker, engine-interop, faction-system, farming-crops, fishing-rod, fps-abilities, fps-armory, fps-deathmatch, fps-network, fps-tactical, fx-ability-vfx, game-endings, hazard-traps, horror-sanity, hud-minimap, hud-mmo, hud-moba, hud-nameplates, hud-open-world, hud-tactical, input-actions, input-pointer-lock, input-remap, input-virtualpad, interaction-feedback, interaction-system, interaction-telekinesis, inventory-equipment, launch-pad, life-ecosystem, life-genetics, light-flashlight, loot-ground, lore-journal, machines-production, match-knockout, mission-system, moba-hero, moba-lanes, mount-ride, movement-boost, movement-dash, movement-flight, movement-grapple, movement-grind, movement-modes, movement-traversal, net-async, net-authoritative, net-claim, net-lobby, net-physics, net-rollback, net-rooms, net-state, net-transport, net-zones, npc-dialogue, npc-pedestrian, party-rounds, photo-mode, present-native, progression-levels, puzzle-circuits, race-kart, render-raytrace, render-realism, research-samples, resource-extraction, resource-nodes, rts-ai, rts-production, run-map, save-online, shared-memory, social-party, spawn-system, sport-golf, stealth-detection, stealth-takedown, storage-containers, survival-needs, tactics-grid, time-rewind, trade-shop, traffic-system, travel-worlds, ui-cursor, ui-endscreen, ui-prompts, ui-theme, units-command, upgrades-boons, vehicle-boostpads, vehicle-damage, vehicle-doors, vehicle-enter-exit, vehicle-rivals, vehicle-spaceship, vehicle-upgrades, wanted-police, weapons-arsenal, weapons-loadout, weapons-mastery, world-access, world-arena, world-atmosphere, world-bindings, world-biomes, world-daynight, world-destructibles, world-district, world-elements, world-fluids, world-grass, world-gravity-zones, world-instances, world-interiors, world-magnetism, world-platforms, world-portals, world-racetrack, world-rooms, world-scene-ground, world-scenery, world-sky-bodies, world-stages, world-streaming, world-tracks, world-vision, world-water, world-weather, wrappers-demo, zone-shrink. Read "third-party" and "porting" + "abi" before writing a presentation module, "modding" before writing a mod, "glossary" for the words. A kit's own API reference is gripforge_gamekit_api. 0 credits.

Free parametric 3D masonry, independent of any game: pillar, arch, fountain, stairs, wall, pavement, rock, cliff, ruin, bridge, watchtower, shrine or a stone guardian. inspect (default) returns real generated bounds, triangle count, metric recipe and assembly sockets. build saves a private textured GLB in the Library, with bevelled geometry, baked vertex shading and named stone/recess/trim/water materials. Dimensions are nominal metres; bounds report actual moulding overhang. Ground origin, stairs ascend toward -Z, <=23 cm risers; top/bottom sockets support assembly. Seed and palette are repeatable. Does not place assets, generate arbitrary AI models, publish, or guarantee navigation/collisions in a target game. 0 credits.

Inspect or create mouth controls on an OWNED GLB while preserving its body skeleton and animations. action=schema returns the contract. inspect returns a persistent job with mesh node indices, local bounds, existing facial controls and compatible dialogue.voice mappings. build automatically reuses recognized Oculus/ARKit/Rhubarb controls, or creates eight draft mouth morphs from explicit mouth landmarks (profile). A sealed mouth requires explicit prepare_mouth=true to cut a guided opening on the output copy and add an interior; it interpolates UVs and skin weights, changes face topology and requires visual review. It never guesses landmarks from the body bounds. Poll gripforge_generation_read for the new private GLB, report, voice_rig JSON and Character Studio link. No automatic source replacement or project binding. No TTS/GPU rental; CPU geometry processing, not full facial expression authoring.

Queue a persistent 240 Hz XPBD simulation using GripForge Scene Engine. Read flexible_chain_schema first. recipe version=1: preset whip|rope|tail, duration<=10s, fps=30|60, physics {length,segments,damping,bendCompliance,gravity,radius,floor,iterations}; choose motion [{time,position,direction,capsules}] OR source {asset:{assetId,revisionId},clip,mount_node,offset,axis,colliders}. Source assets are immutable and workspace-scoped. Returns a job ID; poll generation_read for a new draft skinned-chain GLB, sampled joints, recipe, physics metrics and Character Studio link. This is a neutral physics preview, not a Meshy weapon redesign or automatic rebind of the source weapon. No runtime world collision or seamless loop guarantee. Cancel/retry keeps completed steps. Does not replace or visually validate an existing asset. 0 credits; storage quota applies.

Read the bounded recipe schema for the shared FlexibleChain solver. Presets: whip, rope, tail. Input can follow an exact named mount in a pinned GLB animation or sampled world-space motion. Includes units, collision capsules, limits and a runnable example.

Read the bounded recipe and concept-first workflow for Meshy + Blender flexible weapons, costs, pinned component reuse and source animation requirements.

Google-font CSS + exact HUD copy for a game look (stylish action: Cinzel D–SSS, kart: Teko, tactical FPS: Barlow Condensed). Put letters in HTML/Text nodes. Do NOT rasterize ranks, km/h, or HUD numbers with an image model. 0 credits.

List the mesh formats and grip styles GripForge supports.

Read the native FPS hand template, recipe defaults, limits and reuse workflow. Geometry is manufactured from an articulated FPS template, independently of full-body character generation.

Queue a persistent FPS work asset. Pistol recipe: {version:1,preset:"pistol",name:"Pistol FPS",tempo:1}, tempo 0.7–1.3; manufactures paired arms and articulated pistol clips plus .blend in Blender. Materials recipe: {version:1,preset:"materials",name:"Tactical gloves",source:{assetId,revisionId},node:"FPS_Hands",materials:{map:{assetId,revisionId},normalMap?:{assetId,revisionId},roughnessMap?:{assetId,revisionId},normalScale:0.35,roughness:1,metalness:0}}. Upload and pin owned textures first; they must match source UVs. Select one exact skinned node; other nodes, geometry, fingers and clips stay intact. Normal is OpenGL +Y; grayscale roughness reads G and requires metalness=0 (skin/fabric). Sources <=32 MB, images PNG/JPEG/WebP <=16 MB and 4096 pixels per side. Materials authors a reusable GLB variant, not new hand anatomy or a new .blend; no provider charge. Returns job_id: poll gripforge_generation_read for private Library ref, report and FPS Studio link. Never automatically validates or promotes. Cancel/retry supported; completed steps survive closing the page.

Reusable FPS rules, map and runtime exports for Godot and web: weapons, reload markers, viewmodel, camera, bots, grenades and bomb rounds. Offline bots. Persist changes with gripforge_kit. 0 credits.

One original square icon from the saved game art direction, delivered as PNG, macOS ICNS and Windows ICO in the Library. get returns the latest complete pack. quote prices generation; run queues it after the quoted cost is accepted. Existing icons are reused for free unless regenerate=true. Poll the returned job with gripforge_generation_read. Use files for direct and Steam builds; Steam store artwork is separate.

Store artistic intent on the existing Game Kit project (game.engine), separately from gameplay: look, world/theme, palette, proportions/materials/lighting notes and device budget. Read first, then set with expected_revision. Unspecified fields are preserved; null clears a field or the profile. Existing assets and rendering are not modified. Use this direction with map wizard.look, VFX look, asset prompts, lighting and UI tools, then inspect gripforge_game_asset_plan and actual-engine captures. 0 credits.

Read asset requirements from enabled Game Kits and game.engine archetypes. Without slot: report missing files/bindings, template defaults, metadata style/theme conflicts, unknown classifications and technical ranking concerns. With slot: search accessible Library/Community candidates for that role and the saved art direction; known visual conflicts are excluded. No purchase, generation or binding. Metadata checks do not replace visual/animation/gameplay review in the actual engine. 0 credits.

Read and configure the existing audio.core of an owned Game Kit project. get returns revision, role catalogue, sources, procedural/silent roles, weapon shot overrides, bus defaults, music states/playlists, event cues and broken references. It is a configuration audit; playback stays unverified until heard in the player. configure merges only supplied roles/weapons/volumes/fade and requires expected_revision from get. dry_run previews the same validated result without saving. Preserves other sounds, custom cues, spatial areas and Rust acoustic geometry. Search Library audio first; never use a sound effect as music. No generation, GPU rental, asset purchase or credit charge. Voice identities use gripforge_voice_profile and dialogue.voice; bake audio.speech through Drama 3, English by default.

One report of everything that stands between a Game Kit project (gkp_…) and a FINISHED game of its genre — run it before telling the user a game is done, and again after each round of fixes. It aggregates, by severity (blocker / major / minor): creatures (slots drawn as a capsule, models without a skeleton, missing movement or action clips, the animation chain in progress), controls (no control scheme, the suggested one, keys shared by two actions in one mode, playable actions without a key), content (missing capabilities, required slots, the game.engine report, the core loop rule: an explicit game tick by genre — content.core_loop:tick, fixed for free by fix=true; the source checks are gripforge_core_loop_check), the last headless playtest (failures, warnings, the keys / models / animations probes, the checks of what the game chose — promised gestures, fullscreen, right button, pointer capture, texts —, capabilities no probe covers), the frame rate measured in a real page (sections.framerate: measured, or unverified — never call the game smooth while it is unverified; under 30 fps it is a major defect), keys shown in game (completeness common.no_key_hints, major: a ui.prompts help bar, build.placement keyHints or ui.menu keyHints turned on, a HUD text that lists keys or mouse buttons — no key on screen outside a tutorial), the default screens (sections.screens: on the KitHost page the host mounts them; a compiled module or a custom page must mount the menu, Options › Controls and Video, fullscreen when offered and the prompts from the bindings through the kits’ web views or adapter — a missing kit or a failing screen is a major defect, and a screen no real page has opened is screens.unverified, a major defect: say "not verified", never done) and completeness: what a finished game of the genre has beyond the literal request (MOBA: distinct heroes with Q/W/E/R, minions, towers, jungle monsters, shop, respawn; FPS: weapon, ammo, reload, crosshair, enemies; platformer: checkpoints, collectibles, end of level; racing: laps, ranking, AI opponents… and for every game a way to win AND to lose, title / pause / end screens, sound, a HUD, touch controls), checked against facts of the project, never a declaration. Each defect carries fix { tool, args, note, auto?, credits? }: call it as given. Returns { verdict: complete | playable_with_defects | incomplete, headline, genre, counts, defects, next (ordered calls), sections, text, play_url }. fix=true first applies what is free and safe (the best free Library model on each capsule — rebind at will —, the free standard clips, the suggested control scheme) and reports it in fixes. playtest="run" plays the game headless first (5–20 s); the default reads the last playtest and says when it is stale. Never refuses a write and never spends: paid fixes (a priced model, a rig, a generation) come back with their price to ask the user about. Aim for verdict complete; if you stop before, tell the user exactly what is left. 0 credits.

Requires the Unity SDK. Currently returns 501. Never runs a shell on the host.

Journey — while composing; gripforge_game_audit includes this report, with the creatures, the controls, the playtest and the definition of done of the genre. Report the capabilities a Game Kit project provides, what is still missing for a goal (a preset id from gripforge_gamekit_search or a comma-separated list of capabilities) and which catalogue kits would fill each gap. The installed kits (gripforge_gamekit_installed) come first; call this BEFORE gripforge_gamekit_install to pick the next kit. 0 credits.

Concept art for a whole game, reused before generated. Use it when a user asks for a game like a title they name: call stage=plan with the game as they wrote it. When the user describes specific items (heroes, weapons, props…), pass them in wishes: they are drawn exactly as asked. The plan lists concepts that already exist for that game (GripForge community catalog or the workspace) and theme assets already in the library: reuse them first. It then prices the missing concepts: one style bible, heroes and bosses one image each (MOBA and hero games: each hero with its ability kit and ability icons, see hero_kits), enemies, NPCs, weapons, vehicles, buildings, props and VFX in sheets of 6 cut automatically, textures in sheets of 2 turned into seamless 1024 px textures. Maps (counts.maps) are environment paintings, one image each: the level is then built and dressed to match them. Show the plan and price to the user; only call stage=build with a budget after they agree. Concepts are private review images; nothing is sent to 3D. The plan also returns ambient_kits: the kits of the theme's atmosphere (weather, lighting, snow tracks) with their config; install them in the game project next to the blueprint's gameplay kits.

Journey — the content of the game, as data (documents, config), never as code. Once written, gripforge_game_audit lists what a game of its genre usually has, as proposals (a quest, a way to win or lose, dialogues…) Content is data, not code: generic systems in code (kits, parameters, hooks); levels, characters, weapons, items, balance, texts, assets and inputs as data read at run time; clips, bones, grips and sizes from each asset's passport, never written in code. Read and write the content of a Game Kit project (gkp_…) — what game.engine builds the game from, no engine code needed. kind: archetype (an entity template: kind prop|npc|…, asset { item: lib_… } bound to content:<id> | { slot }, components claimed by kits — see gripforge_game_engine blocks — and gives { quest, offer, progress?, thanks? } for "this NPC gives this quest"), scene (world { kit, config?, doc? }, entries, instances [{ id, archetype, at, yaw?, place?, set? }], zones, layers, ui), script (gripforge.script/1: { id, on: <instance> | archetype:<id> | scene:<id>, rules: [{ when, if?, then, else?, once? }] } with sugars { "interact": "self" }, { "enter": zone }, { "state": { open: false } }, { "quest": id, "is": "completed" }, { "talked": npc }, engine verbs set_state, play_clip, set_solid, show, hide, enable_script, disable_script, load_scene, end_game, call, set_talk and kit verbs as shorthand { "give_item": "key" }), mission (alias quest), dialogue, npc, item, translations, code_module (what game code registers: verbs, components, services, entity kinds), zone, trigger. action=list (kind optional: counts), get (kind, id), put (kind, doc or docs: upsert, validated against the declaring kit schema), remove (kind, id), validate (dry run of a put: the report it would give), place (scene, archetype or asset lib_…, at, yaw?, id?, place ground|none: the placement on the ground is baked once on the server, the same on every engine), bake (re-bake the placements of a scene after its world changed). Writes return the content report after the change; the report never refuses a write — only a document invalid for its schema is. 0 credits.

The hosted orchestration for creating and continuing GripForge games from the web or terminal, one way among others (a Game Kit project made directly, the user's local project with gripforge game-attach, parts written by the user's AI: all welcome and mixable). Any game is accepted: no character, enemy, level, ending or known genre is required. The server owns planning, fabrication, development, checkpoints, budgets and verification of its own tasks; while one runs, do not run a second copy of it locally. If an existing creation id is supplied, get it and resume that exact project, never create a replacement. Ask which engine (Bevy (the GripForge player), Godot, Unity or Unreal) if unspecified and no engine project is in the folder of the user (project.godot, Unity ProjectSettings, an Unreal .uproject, a Bevy crate: then that engine); do not choose silently. A local project is never exported: once built, gripforge game-attach <gc_id|gkp_id> <folder> brings the game into it, whatever the engine. Creation mode is explicit, never guessed silently: when the user has not said it, ask whether the game is generated at once or built step by step. At once: quote build covers the whole saved queue; show its size and maximum price to the user (the first level is playable early, progress.playable). Step by step: each step is quoted, accepted and validated on its own: quote build for the foundation and the first playable level, then quote develop with a message naming the next step, one step at a time; the total duration or size plays no role. A game asked at once that is too long for one build switches to step by step, and the user is told; it is never refused. create saves a brief without spending. When the user already has their own plan (design documents, chosen kits), offer use_plan instead of the AI plan step: use_plan with expected_revision and plan_document (gripforge.game-plan/1: design = the game-designer proposal schema with origin/basis optional, plus optional title kept as written, modules or preset, controls, world, art_direction) is free, uses no AI and no quote, and reaches the same plan-ready state as an accepted AI plan (design agent user). The server keeps every content choice as written: it refuses only an unreadable or structurally invalid document, naming the exact field path, and returns plan_review { ready, diagnostics, suggestions } (kits that realise the features, missing world or controller, features outside the catalog to implement, items naming an absence, unmeasured quality metrics, a journey without a gameplay state). Show them to the user; apply_plan_suggestions with suggestions = the ids the user chose (apply=true) applies only those, never automatically. quote(operation=plan) never replaces a supplied plan; quote refine reworks it with the AI only on request. quote(operation=plan) estimates AI preparation (player_accounts=true adds player accounts through central.client even to a solo game; off by default, online multiplayer always has it), accept(quote_token) starts its durable job ONLY after user approval of the shown price. get returns a persistent operation log, slots, studio_url and the same project to resume. Planning can select an appropriate preset OR compose real engine-compatible modules. A missing preset never means the genre needs a native engine. composition.modules pins all module versions; composition.tasks lists the custom logic still to implement. New plans return design from a separate game-designer agent: named content lists and their counts, animation clips per actor, user-request versus agent-proposal attribution, assumptions and a machine-readable completion contract. Show this saved scope before assembly; refine changes it. get also returns production_sources. quote(operation=produce, production_source=one source ID) estimates a bounded production batch for any contract content, animation, capability or journey obligation; omit production_source to choose the next unprepared source. After price approval, accept runs decomposition, compatible Library matching, authored source documents and at most one concept board for missing visual assets, cropped into individually traceable review drafts. Audio, animation and executable VFX never become complete through a picture. An unavailable matcher with candidates requires review rather than speculative replacement. production.batches persists tasks, dependencies, acceptance checks, candidates, concepts, text, deferred work and diagnostics. asset_matched is not acquisition or integration. draft_ready is not gameplay verification. Changed design fingerprints retain old batches as stale. Use existing concept review/manufacturing and content tools for integration; Develop can read production batches using creation_production. Never report prepared-source coverage as game completeness. Assembly persists that contract on the project before auditing. A design is not implementation or passing evidence. The quoted planning budget includes this bounded stage; old accepted jobs retain their original behavior. New plans also return world: structure (open_world, semi_open, linear, arena, interior or circuit), deterministic seed, validated kit configurations, level documents and remaining work. Assembly applies this saved layout atomically. Reuse it when continuing in the CLI; a visual-theme change does not imply changing geography. A seed change alone is not a designed level. Show world.remaining, including unbuilt architecture or transitions, before calling the game complete. Assembly is a foundation, not a completed custom game: For contract-backed web games, quote build includes a persistent queue derived from every named source and capability, with isolated authoring sessions and independent final verification; there is no twelve-turn limit for the entire game. Review the quoted maximum before acceptance. Unattempted credit-funded tasks are not charged. development.work records reported_unverified and blocked tasks, retained authoring IDs and pending IDs; these are not release evidence. delivery.status=ready requires the build-specific contract and final checks. For an assembled project, quote operation=develop with an optional message to continue the saved queue under a new approved budget, preserving the same project and completed authoring work. The development result distinguishes agent-declared changes, actual checks, remaining work and unsupported runtime capabilities. Each additional pass needs its own estimate. When the contract requires a HUD, new build/develop quotes expose interface_preparation: they reuse the saved concept or include one bounded HUD concept generation and editable document. Its price is included; accepted older quotes never add it. Read creation_hud during authoring, implement the game data in its presenter and use the existing gripforge_hud runtime artistic review. A concept alone does not complete the HUD. Other asset generation requires its own quoted stage. The hosted development worker currently supports target web. For native targets, keep the shared plan, selections and build, then use the engine delivery/import adapter; native custom-code execution is not implemented by the hosted worker. Inside an already running development task, implement only that assigned task; never recursively create or accept a new creation job. After a build, get also returns creatures (the models bound so that no unit is a capsule) and audit { verdict: complete | playable_with_defects | incomplete, headline, fixed (the free fixes applied: models, standard clips, control scheme), defects [{ severity, message, fix { tool, args, credits? } }], next }: show the user the verdict and what is left, work through next (gripforge_game_audit reads it again), and never present a verdict other than complete as a finished game. quote refine with message updates the same plan after approval, preserves selected assets and reviewed work, and persists the conversation. use_reference with slot_id and an owned image asset_id supplies an uploaded concept for visual review. assets searches workspace/community items by actual role before invented generation details. Generic nature packs may cross world themes while keeping the rendering look. New plans consider available styles before inventing one; selection_reason explains planner choices. other_styles=true reveals alternatives with explicit mismatch/unknown compatibility for manual review; never bind a known look mismatch automatically. Prefer reuse before generation; select chooses one or sets mode=generate with a prompt. For generation: quote concept, obtain approval, accept, poll get, show the concept, approve_concept only after visual user review; quote generate, obtain approval, accept, poll and show the model in its studio. A quote with auto_select_character=true selects the missing main character by default when ready; it remains inspectable and replaceable. Other models, replacement generations and older approved jobs remain drafts: approve_asset only after visual review. Never infer visual approval from technical success. quote build then accept acquires the selected community assets and binds them to a Game Kit prototype; generated artwork and models must appear in the accepted quote. The main playable character is presented first. If no usable matching rigged/animated model is found, offer existing ones (gripforge assets show <lib_id> lists their clips) or its generation. A game without a character simply has no such slot. A slot left without an asset uses the module default or stays empty, and a generated model awaiting review is left out until approved; the answer says so (build_incomplete, character_concept_missing). Show missing_assets before approval; never imply that an incomplete prototype is finished. Native engine projects still need CLI delivery/import; a web project has a play_url. Revision required for edits. Every get, list item and action response carries status: idle | running | failed | cancelled | ready, with status_detail { job_id, operation, job_status, stage, message, actions }. Poll get while status=running; stop on failed or cancelled, show status_detail.message and offer the listed actions (retry, dismiss). On any error or failed job, follow client_guidance: explain, offer retry, step-by-step continuation or making that part in the project, then go on as the user chooses. pending alone never means in progress: a failed job stays pending until retried or dismissed. Cancel/retry preserve completed steps. Forward preparation_resume.resume_token through retry when offered. quote_continuation with geometry_mode=designed then accept_continuation with its quote_token use the same saved continuation as the web. platforms on create records web/mobile/windows/mac independently of the engine. A 2D request is made in 3D with a camera that keeps the 2D feel (warning 2d_not_supported). No automatic spending, invented asset IDs, or unrelated template substitutions. Build the engine of the game, never the game in code. The code holds only generic systems: plugged kits, rules set by parameters, hooks. All content is data loaded at run time: levels and scenes, characters, weapons, items, balance values, texts, assets and inputs, in data files or through modding (the game.engine contract), changeable without recompiling, on every engine (Bevy, Godot, Unity, Unreal) and on a local project. Adding a weapon, a level or an enemy is adding data, not code. Each asset is described by its passport (asset.meta, the mesh passport: skeleton, clips, colliders, LOD, weapon grips, sizes): the game reads clips, bones, grips and sizes from it and never writes a clip name, a bone or a scale in code; changing an asset is changing its passport, not the code (gripforge assets show <lib_id> shows it and says what it lacks). Recommend kits with this in mind: the ones whose config, registries and documents carry the content as data. GripForge never refuses a game action for a business reason: when something blocks, it goes on without it and returns warnings [{ code, message, skipped, howToFix }] (applied=false when nothing could be done). Read every warning, tell the user what was skipped, and follow howToFix or offer it; never treat a warning as a finished result. Only security refusals (sign-in, rights, paths outside the sandbox, payment quotas, data of others) remain errors.

Journey — the build order of the content; its last step is gripforge_game_audit, not "it loads". Content is data, not code: generic systems in code (kits, parameters, hooks); levels, characters, weapons, items, balance, texts, assets and inputs as data read at run time; clips, bones, grips and sizes from each asset's passport, never written in code. Read game.engine of a Game Kit project (gkp_…) — a direction, never a gate. action=start (default): the start pack { version, active, structure { engine, kits, content counts, code }, scene, blocks digest, order (scene → playable → assets → archetypes → story → scripts → ui → validate → deliver → code), next [{ step, tool, args, why }], code (where game code goes per engine), rules } and the full content report. action=blocks: the brick catalogue computed from the manifests — entity kinds, archetype components (kit, schema, example), script verbs (conditions / actions with arguments and shorthand), script sugars (interact, enter, leave, scene, talked, quest, state, gives), quest step types, NPC behaviours, worlds, layers, events, variables; scope=catalog adds the bricks of kits not installed (install <kit>). action=points (game.engine 1.3): what can be plugged into the kits without recoding them — without kit, the compact view of every installed kit (hooks with their mutable fields and whether they cancel, events emitted and listened to, registries, collections and config keys with what is moddable or live, settings, verbs, components, services, completeness counters); with kit, the full extension points of that kit with schemas. Look here before writing anything new: the kit or tool that already does it is the proven way, above all for what makes a game complete (input, camera, combat, saving, HUD, menus, sound). The kits are built to be adapted, nothing in them is fixed: adapt one as deep as the game needs through its extension points — its config (gripforge_gamekit_configure), its documents and registries (gripforge_game_content, gripforge_game_project data), taps on its hooks (documents of the hooks collection), scripts whose rules react to its events and change data — in the project, or packaged as a mod (gripforge_game_mods); write something new only when no kit does it, and then as a kit following the extension contract; a local copy of a kit last, saying why. action=report: { ok, errors, warnings, stats } — an error means that piece is left out of the game (the rest plays), a warning that it loads with something missing; each issue may carry a fix { tool, args }. action=next: the next steps only. A game with everything in code is valid: content first, kits second, code for the rest. 0 credits.

Declare that the GAME serves a capability itself instead of a kit (GameProjectDoc.hostProvides), as the engine host serves physics.character: the user's own Godot / Unity / Unreal / Bevy terrain answers world.ground, its own camera camera.*… With the declaration the kits that require it keep working without a provider kit (none is added, none is removed with its dependents) and remove_kit removes that provider kit ALONE in the same revision. The game must then serve it at run time (GDScript host.provide_service("world.ground", obj), Unity Provide, Unreal ProvideService, C gfk_provide, Rust Host::provide — gripforge_docs topic engine-interop); a kit that asks it before it is served logs a warning naming it. Without add/remove it lists { hostProvides, hostable: { <kit>: capabilities the game could serve in its place } }. remove withdraws a declaration: the default provider kit comes back. dry_run=true previews. Then re-run game-attach in the engine project. 0 credits.

Journey — while building or after: a packaged way to apply the same extension mechanisms (config, documents, hooks, reactions) as one separable unit — a variant to switch on, content added over the game, a game opened to its players' mods (an option of the developer). gripforge_game_audit then plays the modded game. The mods of a Game Kit project (gkp_…), docs/mods-engine.md: a mod (gripforge.mod/1) is data applied over the project before the kits mount — config keys and collection documents the kits declare moddable (configPatch by dotted path, data.<collection> { remove, patch, add, order }, hooks documents that tap decisions, scripts whose rules react to events) — whole or not at all, never code. The kits are built to be adapted without recoding them — interceptable (hooks), registrable (registries), modifiable (config, documents, patches) and reactive (rules on events that change data): that is the normal, recommended way to build a game, written in the project (see gripforge_game_engine points) or packaged as a mod. A mod packages those changes as one separable unit: a variant to switch on or off (a hard mode, a season event) without copying the project, content added over the game, and — an option of the developer — a game opened to its players' mods. action=list (default): the game policy (config.mods of game.engine: enabled, disabled, order, deny), every mod with its status (applied | disabled | refused) and issues ({ level, code, path, message }), the documents each added, the fingerprint the network handshake compares. action=validate { mod }: a dry run — the project with this mod added (or replacing the one with its id), applied as the host would: its status and issues, nothing written. action=install { mod }: writes it to the mods collection (validated against the mod schema). action=remove { id }. action=enable / disable { id }: the policy list mods.disabled. action=order { ids }: these mods first, in this order (a mod always comes after those it depends on). 0 credits.

Journey — to look at the running game. Share it as finished only when gripforge_game_audit says complete and gripforge_validate_release says ready; otherwise say what is left. Get the browser play URL of a Game Kit project so a human (or a screenshot step) can run its current revision in the web runtime. Reads the project bundle and returns { url, absolute }; absolute is built on the GripForge origin of this key. Call it AFTER gripforge_gamekit_install or gripforge_game_project bind to check the result live. 0 credits.

Journey — create (or reuse the user's project), then bind a model to each creature slot the game has (get carries the creatures report) and set the controls: action=controls applies the control scheme of the gameplay, a proposal to adapt. Before telling the user the game is done: gripforge_game_audit Content is data, not code: generic systems in code (kits, parameters, hooks); levels, characters, weapons, items, balance, texts, assets and inputs as data read at run time; clips, bones, grips and sizes from each asset's passport, never written in code. Manage Game Kit projects (gkp_…): the container holding installed kits, their lockfile, asset bindings and data collections. action=list lists the workspace projects; create needs name and takes a preset (see gripforge_gamekit_search; adventure is the reference game of game.engine) or an explicit kits map — every new project gets game.engine first (the engine of the game: scenes, archetypes, scripts, life cycle; its answer carries the engine start pack { structure, scene, blocks, order, next, rules } and added, the kits installed for dependencies), then the default kits ui.menu (title / pause / options / load and save screen), input.remap (key remapping), settings.graphics (graphics options), ui.prompts (on-screen key prompts that follow remapping and the gamepad), ui.endscreen (end screen and credits), save.persistence (save slots, quick save, autosave), audio.core (game audio: mixer, music, effects, event → sound table), i18n.text (game texts in the player’s language: translation tables in the translations collection, ICU plurals) ui.theme (the design of every interface as kit config: colours, font, shapes, key glyphs; unset = built-in look) and input.virtualpad (the on-screen gamepad of touch screens, mode auto: shown on a touch screen without a gamepad, Touch controls tab) and display.fullscreen (the game in fullscreen: F11, the Options entry, start in fullscreen with atStart, the player’s choice saved), asset.preload (the files of the game loaded in phases — first playable frame, background, next scene ahead — each file once, checked; progress as state and events shown by ui.menu, ui.endscreen and hud.action) and asset.cache (the files kept on the device between visits: IndexedDB, then Cache Storage with a service worker, then the HTTP cache; quota and eviction) and, on web, ai.navigation (the paths of the game over its real collision geometry: route and steer for NPCs, enemies and click-to-move), addedBy default: keep and configure them (never recode a pathfinding: ask ai.navigation and adapt its config, the hook navigation.plan and the event navigation.blocked; never recode an asset loader, a download queue or a cache: adapt their rules, asset_preload documents, scenes ahead and cache quota instead), leave one out ({ "ui.menu": false }) or remove it only if the user asks. They hold for EVERY game, a compiled module with its own page or a custom page included: that page mounts the menu, Options › Controls and Video, fullscreen and the prompts from the bindings through the kits’ web views (gripforge-ui.mjs, GET /api/v1/hud) or their adapter, never its own copy; get returns the full ProjectDetail (installed kits, bindings, capabilities, missing, playUrl); bind maps asset slots to Library items (lib_…, null to clear); data reads a collection or upserts / removes documents validated against the kits’ schemas (replace=true swaps the whole collection); collection asset_passports = the game’s own passport of a Library model, over the global GripForge one (what it leaves out is inherited): { id: <slot>, item: lib_…, passport: { kits: { lod: { ratios, minTriangles, fit, maxError… } } } } changes the mesh levels of that model for this game only (its own game copy; ratios: [] = no level; characters keep their bone weights); controls reads or sets the controls: without scheme it returns the control schemes of the catalogue (fps, third_person, moba_click, top_down, platformer, fighting, vehicle, rts, point_click: movement by keys or by click, what the mouse is for, expected actions and keys, touch layout), the project scheme, the one its kits suggest, every action with its effective key and the keys two actions share in one mode (warnings with a proposed reassignment, never a refusal — also in get → report.controls); with scheme and dry_run=true it returns the diff of keys the scheme would write; with scheme alone it applies it (input.actions overrides marked source: scheme, the touch layout of input.virtualpad, the actions of the ui.prompts help bar — the bar itself stays off: no key on screen outside a tutorial). Choose the scheme from the kind of game, never from a kit default: Kits offer capabilities that can be switched off and tuned; the game makes one high-level gameplay choice and enables only the coherent set that choice needs: what is not chosen is off and invisible, and no kit default is imposed on the game. The generic modules propose choices, the request picks one: name the kind of game, take its control scheme from the table below (usually the only one allowed), and embed only what that scheme uses on each device — its input sources, its on-screen controls on a touch screen, its settings (remapping, sensitivity, inversion, capture, dead zone only where the scheme uses them). Every option is a choice of the request too: the network input frame is embedded only for a multiplayer game, each other option only when asked. Write the choice into the project: gripforge_game_project create with controls=<scheme>, or action=controls scheme=<scheme> on an existing game (dry_run first, reset=true to drop old key overrides). A dungeon crawler is played with the mouse, not on W A S D. In play, input.scheme (a default kit) implements that choice: it follows the device in hand (keyboard and mouse, gamepad, touch) with the variant of the scheme for it, takes the vehicle scheme while the player drives, gives the input to the interface while a menu, a dialogue or a text field is open, and the Options show per device only the controls the scheme uses; tune it within the choice (input.scheme allowed: the closed list the player may switch between, usually one; variants; composites for held / double / charged / chorded / combo actions), never by enabling every option. Gameplay choices (kind → control scheme, per device): - MOBA → `moba_click` (mouse+keys: click to move, right click = context, settings: remap; gamepad: stick, aim assist, settings: remap, deadzone, aim_assist; touch: tap to move, on screen: tap_world, settings: touch_layout). A hero is ordered with the mouse (right click to move / attack, abilities toward the cursor); a gamepad moves it with the stick. Interaction feedback: interaction.feedback style bold, on: highlight, cursor, confirm, fail, sound, orderMarker; interaction.system pointerTarget: true. - Real-time strategy → `rts` (mouse+keys: click to move, right click = context, settings: remap; gamepad: click to move, settings: remap, deadzone; touch: tap to move, on screen: stick + buttons + tap_world, settings: touch_layout). Units are selected and ordered with the cursor; nothing is moved with W A S D. Interaction feedback: interaction.feedback style default, on: highlight, cursor, confirm, fail, sound, moveMarker, orderMarker; interaction.system pointerTarget: true. - Dungeon crawler / action RPG seen from above → `arpg_click` (mouse+keys: click to move, right click = context, settings: remap; gamepad: stick, aim assist, settings: remap, deadzone, aim_assist; touch: tap to move, on screen: buttons + tap_world, settings: touch_layout). Played with the mouse: click to walk, click an enemy to attack, right click for the special; the left stick on a gamepad, a tap on a touch screen. W A S D is not the default for this genre. Interaction feedback: interaction.feedback style default, on: highlight, nearbyHint, worldPrompt, cursor, holdGauge, confirm, fail, sound, orderMarker; interaction.system pointerTarget: true. - Top-down shooter / twin-stick roguelike → `top_down` (mouse+keys: WASD, right click = aim, settings: remap; gamepad: stick, aim assist, settings: remap, deadzone, aim_assist; touch: virtual stick, on screen: stick + buttons, settings: touch_layout). Constant dodging needs direct movement (W A S D / left stick) with the cursor or right stick to aim. The game may also allow: arpg_click. Interaction feedback: interaction.feedback style default, on: highlight, worldPrompt, holdGauge, confirm, fail, sound. - Point-and-click adventure → `point_click` (mouse+keys: click to move, right click = context, settings: remap; gamepad: stick, aim assist, settings: remap, deadzone, aim_assist; touch: tap to move, on screen: buttons + tap_world, settings: touch_layout). A free cursor: click where to go and what to use. Interaction feedback: interaction.feedback style default, on: highlight, nearbyHint, worldPrompt, cursor, confirm, fail, sound, moveMarker; interaction.system pointerTarget: true. - First-person shooter / survival → `fps` (mouse+keys: WASD, captured mouse, right click = aim, settings: remap, mouse_sensitivity, invert_y, pointer_capture; gamepad: stick, right stick looks, aim assist, settings: remap, deadzone, stick_sensitivity, invert_y, aim_assist; touch: virtual stick, right stick looks, on screen: stick + buttons, settings: touch_layout, stick_sensitivity, invert_y). W A S D and the captured mouse to look (right stick on a gamepad). Interaction feedback: interaction.feedback style subtle, on: highlight, worldPrompt, holdGauge, confirm, fail, sound. - Third-person action / open world → `third_person` (mouse+keys: WASD, captured mouse, right click = aim, settings: remap, mouse_sensitivity, invert_y, pointer_capture; gamepad: stick, right stick looks, aim assist, settings: remap, deadzone, stick_sensitivity, invert_y, aim_assist; touch: virtual stick, right stick looks, on screen: stick + buttons, settings: touch_layout, stick_sensitivity, invert_y). W A S D relative to the camera and the captured mouse (right stick) to turn it. Interaction feedback: interaction.feedback style default, on: highlight, nearbyHint, worldPrompt, holdGauge, confirm, fail, sound. - Platformer → `platformer` (mouse+keys: WASD, settings: remap; gamepad: stick, settings: remap, deadzone; touch: virtual stick, on screen: stick + buttons, settings: touch_layout). Direct movement and jump, no pointer. Interaction feedback: interaction.feedback style subtle, on: worldPrompt, confirm, fail, sound. - Fighting → `fighting` (mouse+keys: WASD, settings: remap; gamepad: stick, settings: remap, deadzone; touch: virtual stick, on screen: buttons, settings: touch_layout). One key per blow, no pointer. No interaction.feedback. - Racing / vehicles → `vehicle` (mouse+keys: vehicle, settings: remap; gamepad: vehicle, settings: remap, deadzone; touch: vehicle, on screen: stick + buttons, settings: touch_layout). Throttle, brake and steering; no pointer. Interaction feedback: interaction.feedback style subtle, on: worldPrompt, confirm, fail. - Turn-based combat / tactics → `point_click` (mouse+keys: click to move, right click = context, settings: remap; gamepad: stick, aim assist, settings: remap, deadzone, aim_assist; touch: tap to move, on screen: buttons + tap_world, settings: touch_layout). Each turn is played with the cursor or a tap (combat.turns, tactics.grid): pick a unit or an action, then its cell or its target; no direct movement. Interaction feedback: interaction.feedback style default, on: highlight, cursor, confirm, fail, sound; interaction.system pointerTarget: true. - Card game / deckbuilder / pack opening → `point_click` (mouse+keys: click to move, right click = context, settings: remap; gamepad: stick, aim assist, settings: remap, deadzone, aim_assist; touch: tap to move, on screen: buttons + tap_world, settings: touch_layout). Cards and packs are played with the cursor or a tap (cards.deck on combat.turns; collection.packs for a pack-opening or gacha game with an album): pick a card or a pack, then its target or the next object to turn over; no direct movement. Interaction feedback: interaction.feedback style default, on: highlight, cursor, confirm, fail, sound; interaction.system pointerTarget: true. - Puzzle → `third_person` (mouse+keys: WASD, captured mouse, right click = aim, settings: remap, mouse_sensitivity, invert_y, pointer_capture; gamepad: stick, right stick looks, aim assist, settings: remap, deadzone, stick_sensitivity, invert_y, aim_assist; touch: virtual stick, right stick looks, on screen: stick + buttons, settings: touch_layout, stick_sensitivity, invert_y). The puzzle kits (puzzle.circuits, puzzle.beams) sit in the world: the player reaches them with the game's character, seen from behind by default, in first person or with the cursor when the game is built that way. The game may also allow: fps, point_click. Interaction feedback: interaction.feedback style default, on: highlight, nearbyHint, worldPrompt, holdGauge, confirm, fail, sound. - Sports / ball game → `third_person` (mouse+keys: WASD, captured mouse, right click = aim, settings: remap, mouse_sensitivity, invert_y, pointer_capture; gamepad: stick, right stick looks, aim assist, settings: remap, deadzone, stick_sensitivity, invert_y, aim_assist; touch: virtual stick, right stick looks, on screen: stick + buttons, settings: touch_layout, stick_sensitivity, invert_y). sport.ball on foot: the athlete moves with W A S D or the left stick and kicks where it faces; in its car mode, the vehicle controls. The game may also allow: vehicle. Interaction feedback: interaction.feedback style subtle, on: worldPrompt, confirm, fail, sound. - Golf / mini-golf → `point_click` (mouse+keys: click to move, right click = context, settings: remap; gamepad: stick, aim assist, settings: remap, deadzone, aim_assist; touch: tap to move, on screen: buttons + tap_world, settings: touch_layout). sport.golf: the shot is aimed with the free cursor (the ball goes where the pointer is) and played with a click or a tap that stops the power gauge; nobody walks, so no W A S D and no captured pointer; the left stick turns the aim on a gamepad. No interaction.feedback. - Flight / space combat → `vehicle` (mouse+keys: vehicle, settings: remap; gamepad: vehicle, settings: remap, deadzone; touch: vehicle, on screen: stick + buttons, settings: touch_layout). Throttle, yaw and pitch on the keys or the sticks, no pointer: vehicle.spaceship pilots a starfighter (the space_combat preset) and the same vehicle controls drive a plane. Interaction feedback: interaction.feedback style subtle, on: worldPrompt, confirm, fail. Per device: the cursor per kind of target only with a mouse; on a gamepad the prompt shows the button of the active gamepad, on a touch screen the text alone (ui.prompts / input.actions promptFor); the designated target comes from input.scheme target() (cursor, aim axis, stick). Interactions must be visible. Install interaction.feedback for any game with interactive objects (not for a pure fighting game) and set its elements from the gameplay table row; never draw an interaction prompt in a kit or a script: kits that offer an interaction pass their kind to interaction.use.register(entity, label, action, kind) (or interactable.kind on an archetype), and slow uses get a hold (interactable.hold / interaction.system.holdSeconds) instead of a custom timer. Styles and kinds are data (feedback_styles, interaction_kinds): restyle with them, do not recode. Leave the player settings (interaction.feedback.settings) to the menu. A scheme is a starting proposal: a preset applies its own at creation (create takes controls to pick another one — pass the scheme of the kind of game the user asked for — or none), and any key stays changeable with gripforge_gamekit_configure input.actions { overrides: [{ action, keys, buttons?, disabled? }] } — the action of any kit; those entries always win over the scheme. logic chooses where the kits’ logic runs: wasm = logic of the kits in Rust compiled to wasm (Bevy player), the only logic; default wasm (create takes logic too; action=logic without logic reads it, with logic sets it). A game installing a kit without a Rust module cannot start: logic reports runs: none and fallback: the kits. delete needs confirm=true. Call this first to create or pick the project, then gripforge_gamekit_installed (existing project) or gripforge_gamekit_install. 0 credits.

Publish the Game Kit project so its play link opens without an account, and start its online server if the game is online (same API as the Publish button and `gripforge game-publish`). mode=preview shares a TEST build: no release checks (completion, animation proof), the project is marked preview; use it to let testers join an online game before it is release-ready. mode=release (default) runs the release checks and refuses with the blockers if the build is not ready. unpublish=true takes the game and its online server offline (mode is ignored). Needs an editor role on the owning workspace. Returns published, publicationMode and playUrl. 0 credits.

Read immutable game release history and stable/beta channel revisions; obtain a snapshot fingerprint for an exact saved project revision; register a prepared signed release; or promote its build ID with expected_revision from history. Direct Windows/macOS updates use the signed launcher feed. Steam builds use SteamPipe and are promoted in Steamworks separately. Registering never promotes. Use gripforge game-release prepare/finalize/publish locally for packaging, signatures and binary uploads; game-release steam-upload uses an already authorized local SteamCMD account. Never send a signing private key or Steam password to this tool. A missing artifact returns warnings and preserves the current channel.

The scenes and levels of a Game Kit project (gkp_…), game.engine 1.7 — use this tool, never code a scene manager. A scene has a kind: interface (no world: title, main menu, credits — kits build nothing, the renderers draw no world), 2d, 3d (default), loading (shown while the next scene loads), cinematic; a role: menu (entered first, Play leads on), level (levels follow the order), credits, other; next (the scene that follows: Play from the menu, the end of a level, the script verb next_scene); loading (how it shows when its files are not ready — on demand: preloaded or cached files go straight in, a screen comes up only after sceneFlow.showAfterMs; overlay = ui.menu's loading screen with the real asset.preload progress, none, background = the scene left stays until this one is ready, or a scene id: a loading scene or an interface scene of the game); loadingScreen { image, tips, background, title } (several screens per game: one per scene, or flow.loadingScreen); background { image, video, color, scene } of an interface scene; unload (release | keep); transition (replace: the scene left is unloaded | add: it stays loaded under this one) and below (what the scenes under an added scene do: run, pause = drawn but frozen, hide = not drawn). Scenes stack and keep their states at run time — the game changes scene by program (engine.scenes.enter(scene, entry, { unload: false, previous: { visible, paused } }), push / pop / unload / activate / set(scene, { visible, paused }); script verbs load_scene { unload: false }, push_scene, pop_scene, set_scene): a scene kept loaded is frozen in memory and resumed at once; nothing a loaded scene uses is freed. The flow (game.engine config sceneFlow): transition replace | add, menu scene | overlay (scene: the title of ui.menu is a scene of its own, entered first with no world behind it — synthesised as menu when the project writes none; every new game and every opened game has it), menuScene, loading overlay | scene | none | background, loadingScene, showAfterMs, minLoadingMs, loadTimeout (s), onFail retry | enter | wait, retries, loadingScreen, unload. Each scene loads its files before it is entered (asset.preload, progress on the loading screen) and leaving it frees its instances and the memory of the files only it used (never the disk cache); hooks engine.scene.change (redirect or refuse a change) and engine.scene.release, events engine.scene.loading / loaded / unloaded / refused for mods. The asset list of a scene (scene_assets: files with role, phase, size, shared) is generated from its content (instances → archetype models, the global slots a played scene needs, world, HUD, music, menu art) — retouch it by hand or with assets_put; asset.preload uses it as one more source, never a restriction. action=list (default; scene: one scene with its files and plan): the flow (menu, start, levels, loading), each scene (kind, role, next, instances, world kit, files, written or generated list), the scene issues of the report, the check of the asset lists and the load plan (per scene in play order: what entering it loads, what leaving it frees, what is shared, and the loading screen of that transition). action=create { id, kind?, role?, title?, next?, loading?, background?, world?, at? (index) } | update { id, set: {…} } (null clears a key) | remove { id } | order { ids } (these first, in this order) | link { id, next } (next null: unlink) | flow { set: { menu, menuScene, transition, loading, loadingScene, showAfterMs, minLoadingMs, loadTimeout, onFail, retries, loadingScreen, unload } } | menu { id?, title?, background? } (writes the menu scene as a document, first, role menu) | assets { scene? } (generate or refresh the lists; files written by hand are kept) | assets_put { scene, files } | assets_check (missing file, file used by nothing, duplicate, list out of date). dry_run answers the new state without writing. Every write answers the new state. Scene content (instances, entries, zones, scripts) stays with gripforge_game_content. 0 credits.

Journey — after each change: it also checks what the game chose (its scheme's gestures, fullscreen, right button, pointer capture, texts against the bindings). A playtest that passes is not a finished game: gripforge_game_audit (playtest="run" runs this test) gives the verdict and everything the tester cannot see. Play the project headless and report what works and what is broken. Run it after building or changing a game, before telling the user it is done. An automatic player spawns, walks, follows the mission waypoints, uses what offers a prompt, picks things up and attacks enemies. Probes check each capability the game provides (character control, combat, XP and levels, inventory, interactions, missions, needs, gathering, crafting, building, machines, world counters). Detectors flag a stuck player, unreachable goals, kit exceptions, NaN positions, a player under the ground, enemies that never reach the player, and step time. Then it checks the game on what it CHOSE, on a fresh run (lines game.*): the gestures its control scheme promises (a click game: click to walk, held to follow the cursor, click an enemy to strike, held to keep striking), fullscreen when the game offers it, the right button when the game gives it a role (no browser menu), the pointer captured exactly when the scheme looks with the mouse, and its texts (title, help, tutorial, dialogue) against the real bindings — never a fixed W A S D in a click game; nothing is imposed on a game that did not choose it. The frame rate is · not verified here: only a real page measures it (the play page reports it to the audit). Returns { summary, text, report, next }: text has one line per result, ✓ pass, ⚠ warning, ✕ failure, · not checked, each with its time, position, entity or kit; "provoked" means the tester set the condition up through the kit service. Fix the ✕ lines first, then run it again. report.notCovered lists the capabilities no probe looks at. report.replay replays the same inputs; pass it back as replay to check a fix against the same run. Deterministic: same project, seed and budget, same report. Runs in the compute service, about 5–20 s. 0 credits.

Generic concept → theme pipeline for a Game Kit project: the saved art direction (gripforge_game_art_direction: look, theme, palette, notes, device) becomes ONE interface concept image (title, main menu, HUD with health, a weapon card, key prompts), a vision pass reads its design tokens (colour roles, radius, border width, button shape, padding, letter case, weight, the closest open-licence font, glyph style, closest preset) and they are written as the ui.theme config (contrast repaired). The concept is kept in the Library (kind concept, category ui). action=quote prices it: 1 credit per run (quote it and ask the user before run). Without regenerate the latest concept of the project is reused with the tokens already read from it (0 credits), so a run after a dry run applies exactly what the dry run showed; reread=true reads the concept again (1 credit). dry_run=true writes nothing. run returns a generation job: poll it with gripforge_generation_read; its result carries config, tokens, fixes, previous_config and revision_before for a rollback (gripforge_gamekit_configure ui.theme with previous_config).

Journey — last, and never for a local engine project (gripforge game-attach works in it). Run gripforge_game_audit then gripforge_validate_release first: export a game whose verdict is complete and whose release is ready, or tell the user what is left. Not for a local engine project: when the game's Godot project already exists on the user's machine, do not export it — attach GripForge to it with the CLI, gripforge game-attach <project_id> <project folder> (addon gripforge_kits + game under gripforge/, never its project.godot, scenes or scripts); Unity / Unreal: gripforge_gamekit_deliver in place. A native project's answer carries notices [{ code: export_not_needed, howToFix }]: follow it. If game-attach reports manual_setup_found, follow its howToFix before any --force. Prepare a download manifest for an existing Game Kit project in the current workspace. Reuses the same player and HUD as hosted play, compiles only installed enabled kit runtimes, and lists the game configuration, static resources and Library assets to download. Does not publish the project or change its visibility. The hosted MCP cannot write local files: use gripforge game-export <project_id> <folder> locally, then npm run dev. The CLI verifies downloads and preserves local edits on updates (--check previews conflicts; --force backs up and replaces edits only when authorized). The delivered versions remain fixed until an explicit export update. Legacy kits and unpublished source overlays cannot be exported. net.* kits still need their online services. 0 credits.

Read source-linked plans (get), author one (plan), construct actual kit content atomically (build, with plan_id, options and expected_revision; dry_run previews; activate explicitly sets the entry scene), save an input journey (journey), inspect without a run (inspect), or queue a real Bevy input/capture test (test). get section:builder with kit returns the provider options; section:journey returns its schema. Builders belong to kits, never genres. test returns a durable job; AI visual review uses a separately quoted Develop and test pass. Generated content remains unverified. Preserve requirements and existing authored content. 0 credits.

Read every spatial obligation in the saved game contract (get), or save region/location classifications, hierarchy, entries and intended connections (put). No genre counts. Old contracts need explicit classifications; declared spatial roles cannot be weakened. Regions require real child locations. Construction and verification remain separate. Scope writes use project revision checking. Continue with gripforge_game_world build/journey/test. 0 credits.

The animation layers of the characters of a Game Kit project (gkp_…), part of the generic animation layer of the kits (module-kits packages/gamekits/src/runtime/anim-layers.ts, mixed in Rust by the players: Bevy on the web and natively, libgfkits for Godot / Unity / Unreal). An action (attack, heavy_attack, punch, kick, ability, shoot, reload, throw, block, emote, interact, hit, body) plays on the parts of the body its layer takes over the locomotion, which keeps the rest: attacking while running keeps the legs running and the player moving. Defaults: a one-hand attack while moving = upper body (legs and hips to the locomotion, waist blended), standing still = whole body; movement free. NEVER recode a blend of animations, a bone mask or a movement lock in game code: set the layer here. action=table (default): the masks (full, upper, torso, lower, legs, arms, arm_right, arm_left, head, upper_right, upper_left; regions hips, spine, chest, head, arm_l, arm_r, leg_l, leg_r), the action families with their states and their default layer. action=get: the layers the project sets — actionLayers of its character kits (character.third_person, character.first_person, character.top_down: by family or by state such as attack3) and weaponLayers of weapons.loadout (by weapon id). action=set { kit, key, layer }: one entry, layer = { mask, still, weight, mode: replace | additive, fadeIn, fadeOut, movement: free | slow | lock | root, moveScale } (mask / still: a mask name or { regions: { <region>: 0..1 }, bones: { <bone>: 0..1 } }); the kit validates it, mods may override it (the keys are moddable), the hook character.action decides the movement once per action and player.action reports it. action=clear { kit, key }: back to the defaults. Existing games get the defaults without any change; nothing is recompiled. 0 credits.

The API reference of one Game Kit, as compact Markdown: what it does and when to use it, every config key (type, default, unit, limits; mod = a mod may set it, live = read again while the game runs, scene = a scene may override it), hooks (payload, mutable fields, cancellable), events emitted and listened to (when, payload), web views and HUD vars, services and their methods, script verbs, asset slots, data collections, input actions, mod points, its Bevy presentation module (needs, ABI families, web UI), target parity, changelog, and an example config and an example mod (gripforge.mod/1). Read it BEFORE gripforge_gamekit_configure, before writing a mod (gripforge_game_mods) or hooks / scripts for the kit, and before writing its presentation module. sections narrows the answer (e.g. ["config","hooks"]); full=true adds nested keys, examples of every hook and verb, and the README. 0 credits.

Journey — while composing, and to fix what gripforge_game_audit reports (its defects carry the exact config to write). Content is data, not code: generic systems in code (kits, parameters, hooks); levels, characters, weapons, items, balance, texts, assets and inputs as data read at run time; clips, bones, grips and sizes from each asset's passport, never written in code. Tune an installed kit: merge config values validated against the kit configSchema (read it with gripforge_gamekit_get), or patch the config as a document (patch: set a nested key, add / change / remove one element of a list without rewriting it), or toggle enabled to switch the kit off without uninstalling it. Configure installed kits freely rather than recoding what they do (game.engine: startScene, startEntry, flow.title, story flags). The interface look of a game is ui.theme: set preset (scifi-lowpoly, military, fantasy-medieval, cyberpunk-neon, cartoon-bright, horror-dark, minimal-clean, retro-pixel) for a complete look, and only the colours, typography, shapes or glyphs that must differ (they win over the preset); gripforge_game_theme_from_concept fills it from the art direction. No key on screen in game outside a tutorial: never write a legend of the controls into a HUD text (hud.action hint, captions, labels) and leave ui.prompts bar off (its default); a tutorial step pushes its prompt with ui.prompts show(), and ui.prompts context: true / interaction.feedback worldPrompt: true are set only when the gameplay needs contextual prompts (the key is then the live binding, never written) — the controls live in Options › Controls (input.remap). Switching game.engine off follows the confirmation of a removal: the first call answers 409 confirm_required with details.confirm; send it back with enabled=false after the user agreed — never on your own. Returns { kit, project } with the new project revision. Call this AFTER gripforge_gamekit_install. 0 credits.

Journey — when the game has creatures, right after creating or binding: no unit stays a capsule, every model gets its movement and action clips (read from its passport, never named in code). gripforge_game_audit folds this report with the rest and applies the free steps with fix=true The living units of a Game Kit project (gkp_…) — player, heroes, enemies, bosses, minions, monsters, NPCs: the creature slots its kits declare. Two rules: NO creature ships as a capsule, and a model ships WITH its animations (movement AND action). action=report (default): every creature slot with its status (bound | fallback = a capsule in game), the model bound, the clips the slot plays (locomotion, action), the ones the model lacks, and its chain (steps: clone | rig | animate | more_clips, each done / running / proposed with its tool, free or its credits). action=propose: ranked Library models for each unbound slot — the workspace first, then the Community — with the reasons of the rank (source, slot and role words, game theme, size, rigged, animated, triangle budget). YOU judge: read the reasons, pick the model that fits the game (any monster beats a capsule), bind it with gripforge_game_project { action: "bind" }; nothing is final, rebind any time. A slot with no proposal at all carries defect creature_no_model: generate a model (paid — quote and ask first), never leave the capsule. action=animate: run the free steps of the chain now for the bound models (a free Community model is copied into the workspace and bound in place of the original; the standard clip pack is retargeted, 0 credits; wait=true waits for the result, else it runs in the background — read report again). action=fill: bind the best free proposal to every unbound creature slot, then animate (what the hosted creation does). Nothing here spends credits: the automatic rig, a priced Community asset and generated motions always come back as proposed steps for you to quote and ask about. Binding a model through gripforge_game_project already starts its chain. 0 credits.

Journey — kits into a local engine project, in place (with the CLI: gripforge game-attach <gkp_id> <folder>, Bevy, Godot, Unity or Unreal; never an export); a final delivery is last. Run gripforge_game_audit then gripforge_validate_release first: deliver a game whose verdict is complete and whose release is ready, or tell the user what is left. Content is data, not code: generic systems in code (kits, parameters, hooks); levels, characters, weapons, items, balance, texts, assets and inputs as data read at run time; clips, bones, grips and sizes from each asset's passport, never written in code. Turn catalogue kits, or the enabled kits of a Game Kit project, into engine files: Godot gets GDScript, scenes and config under gripforge/kits/<kit>/, Unity gets an embedded UPM package plus C# files and Unreal a Blueprint-first plugin for every kit with an engine delivery (gameplay and net.* kits; recipes otherwise), web needs nothing. Kits always deliver at their latest version; a kit the project copied locally is delivered from its copy. Call with dry_run=true FIRST: it returns the plan (files to add / keep / modify / delete, conflicts, dependencies, post-install steps) and never charges. Pass project_state (the current gripforge/gamekits.lock.json, path → sha256 of the files under gripforge/, engineVersion) so updates keep your edits and detect conflicts. Then call without dry_run to get the bundle. The hosted endpoint cannot write into your project: write bundle.files following plan.actions in order (backup, write, writeBin from bundle.binaries, delete with .uid / .import siblings, writeLock with bundle.lock), or use the npm client with project_dir (gripforge_gamekit_deliver_local), or the editor bridge. A blocked plan answers 409 with the plan: a file edited in the engine project is reported (modified_file), never overwritten silently; force backs it up then overwrites it. Credits: the first delivery of a kit major version to an engine costs 1 credit per workspace; re-deliveries, updates within a major, dry runs and the web target are free.

Dependency graph of a Game Kit project: nodes (installed kits with version, provides, requires), edges labelled by capability, same-scene overlaps (conflicts: several enabled providers of one exclusive capability) and unresolved requirements. Pass project_id for the installed graph; pass a kit id instead to read the declared requires / provides tree of a catalogue kit before installing it. Call this BEFORE gripforge_gamekit_remove with force. 0 credits.

Read one Game Kit: its manifest (provides / requires capabilities, asset slots, data collections, events), README docs, config JSON schema with defaults, version and changelog (information only: projects always run the latest version), and summary { tagline, tags, targets, status, cover, coverUrl } with the featured image URL. Pass with_usage=true to also list the workspace projects that have it installed. Call this BEFORE gripforge_gamekit_configure to know which config keys exist. 0 credits.

Journey — while composing. A kit installed and left empty is not done (audio.core without a sound, ui.endscreen without an end condition): gripforge_game_audit names them. Content is data, not code: generic systems in code (kits, parameters, hooks); levels, characters, weapons, items, balance, texts, assets and inputs as data read at run time; clips, bones, grips and sizes from each asset's passport, never written in code. Add a gameplay kit (a brick) to a Game Kit project (gkp_…); installing marks that the game uses it. Decide it yourself, no confirmation needed. Always the latest version; dependencies are added automatically and listed in added [{ kit, version, requiredBy, capability }] (e.g. game.engine for a kit that requires it). A kit that looks with the mouse (first person, third person, shooters, vehicles, mounts) also brings its companion input.pointer_lock (mouse capture) the first time it is added — listed in added with reason "companion of …"; keep it unless the user wants the cursor visible (remove it alone with gripforge_gamekit_remove), and never code requestPointerLock by hand. Kits are bricks that go together: two kits handling the same thing (e.g. two lightings) are a same-scene overlap, never a game-level conflict — only one is active in a scene. Returns { plan, applied, added, project }; dry_run=true previews without writing. Check gripforge_gamekit_installed first, then gripforge_gamekit_search; tune with gripforge_gamekit_configure. To adapt a kit to the game, go through its extension points first — its configuration, its overridable collections, rules and scripts on its events, hooks documents on the hooks it declares (engine.hooks) — and recode or copy a kit only as a last resort, saying why: a copy is no longer updated. Legacy genre kits (legacy.*) use gripforge_kit. 0 credits.

Journey — first on an existing project; gripforge_game_audit then says what stands between it and a finished game. Start here for an existing project. Lists the kits installed in a Game Kit project (gkp_…), first and marked priority: id, name, enabled, addedBy (user / agent / auto / preset / default — the interface kits every new game gets), provides, requires and config, plus the rule that goes with them. Versions are for your information (projects always follow the latest): each kit also has version (in the project), latest, delivered { godot | unity | unreal } (the version the last engine delivery carries — a delivered game can run an older one) and pendingMigrations (manual migrations left to you). Installed kits are choices already made for this game: build on them for what they cover and configure them as needed; search the catalogue only for what they do not cover. 0 credits.

Uninstall a kit from a Game Kit project (gkp_…). Remove a kit the user installed (addedBy user in gripforge_gamekit_installed) only when the user asks. game.engine (the engine of the game) and the default kits (ui.menu, input.remap, settings.graphics, ui.prompts, ui.endscreen, save.persistence, audio.core, i18n.text, ui.theme, input.virtualpad, display.fullscreen, asset.preload, asset.cache, ai.navigation on web, addedBy default) are removed only when the user explicitly asks — never to "clean up". A kit other kits depend on, and game.engine always, is not removed by a first call: it answers 409 (kit_in_use, or confirm_required) with details { dependents, confirm } — nothing applied; show the user the dependents, then call again with with_dependents=true and confirm=<token> to remove the kit and its dependents together (the token is bound to the project revision). dry_run=true returns the plan, the dependents and the token without writing. force=true (legacy) removes it and disables its dependents instead; prune=true also drops the dependencies nothing else needs (never game.engine nor a default kit). When the GAME serves what the kit provided (its own terrain for world.ground, its own camera…: a Godot, Unity, Unreal or Bevy project attached with game-attach), pass host_provides=["world.ground"] (or [] for every capability the dependents need from it, listed in details.hostable of the 409): the kit goes ALONE, its dependents stay, and the project declares those capabilities served by the game (hostProvides) — offer this to the user before removing dependents. Returns { applied, removed[], dependents, project }. 0 credits.

Journey — first: analyse the project or the plan, then recommend the kits that fit, each with what it brings, and offer to plug them in ("these kits would help: shall I plug them in?"); never impose them. A preset of a known genre has a `done` list: proposals of what a finished game of that genre usually has, recorded in the contract (gripforge_define_game) when the user takes them; gripforge_game_audit checks what was chosen at the end Content is data, not code: generic systems in code (kits, parameters, hooks); levels, characters, weapons, items, balance, texts, assets and inputs as data read at run time; clips, bones, grips and sizes from each asset's passport, never written in code. Start here. Search the modular Game Kit catalogue (vehicle.driveable, mission.objectives, npc.wanted, …) by free text, capability, tag or engine target. Pass project_id to score kits against that project: each hit then carries state { installed, enabled, fills, reason } and kits that fill one of its missing capabilities rank first. Kits are bricks that go together: a kit handling the same thing as an installed one is a same-scene overlap (only one active in a scene), never a reason to skip it. Returns { kits, presets }; each kit carries its presentation: name, tagline (one-line hook), tags, targets and status per engine (web / godot / unity / unreal), and cover / coverUrl, the featured landscape image (WebP) to show the user when proposing kits. legacy.* kits play through their existing client. Kart Racing is native: compose world.racetrack + vehicle.driveable + race.kart; edit race_tracks through project data. world.lighting adds shared sun/fill, point lights and spots, presets, bounded light budgets and bloom; edit world_lights through project data. world.fluids adds interactive mud / blood surfaces, displaced 3D wheel ruts and footprints; edit fluid_surfaces through project data (web renderer). world.terrain adds a seeded large landscape (terrain_edits data), movement.traversal climbing / swimming / gliding with stamina, world.elements data-driven fire, water, ice, electricity and poison (element_rules, element_materials, element_climates data). Also returns packs: content packs (ready-made game content — documents and kit configs, e.g. a realistic atmosphere, a tree of life, a star system) that gripforge_gamekit_install { pack } installs into a project. Call this BEFORE gripforge_gamekit_install. 0 credits.

Projects always follow the latest version of their kits: the studio updates them automatically when a project is opened. Use this to do it now. Dry-run by default: returns the plan (from / to version, migrations). Pass apply=true to apply it. A migration with auto=false is never applied by itself: it is your task — apply its note (config, data) then call again. Omit id to check every installed kit. 0 credits.

CREATE any mesh asset and save it to Library — the canonical generation door. kind decides the pipeline: character/enemy/boss are auto-rigged (pass concept_item or file_url for image-to-3D; otherwise text-to-3D), weapon and prop are direct text-to-model (no rig, no concept step). The workspace Style DNA is applied to the prompt. If the look may already exist, call gripforge_style_kit FIRST. Costs 10 credits. Returns a persistent job_id immediately; poll gripforge_generation_read for progress and the Library/Studio result. Closing this call does not cancel it.

Reusable plan → artistic concept → isolated reference views → review → Meshy or explicitly selected Tripo build workflow. Generate 1–4 images through OpenAI or Imagine; alternate views reference one master but must be checked for actually different viewpoints. If edits repeat the front, use one good concept instead of a false multiview set. Separate explicit budgets for concepts and 3D. Accepts text, owned concept views or a static GLB. Optional delivery preserves the source master and prepares reusable LOD GLBs with PBR and a common metric fit. Durable jobs return workspace links; poll generation_read, cancel/retry preserves finished steps. Review the actual render before publishing. No guaranteed interiors, collisions or runtime LOD switching; no automatic game replacement.

CREATE a Mixamo T-pose auto-rigged character, enemy or boss and save it to Library. NEVER describe a handheld weapon, staff or shield in the prompt: characters and enemies are forged empty-handed (the forge enforces it); make the weapon apart with gripforge_generate (kind=weapon) and put it in the hand with gripforge_attach. polycount is honoured (default 8000; keep it at or under 10000 for a mobile game).To sculpt from a picture (image-to-3D): pass concept_item (lib_… from gripforge_concept_correct / character_concepts) or file_url. Without an image this is text-to-3D. If the look may already be in the locker, call gripforge_style_kit FIRST. For knives/guns use gripforge_generate_weapon. Alias of gripforge_generate (the canonical door). Costs 10 credits. Returns a persistent job_id immediately; poll gripforge_generation_read. Closing this call does not cancel it.

Generic animal/alien creature factory, one to eight original creatures per batch. stage=plan quotes concept + Meshy 3D + anatomy vision without paid calls. stage=build requires explicit budget and stable idempotency_key, then generates private meshes and inspectable skeleton overlays. Poll gripforge_generation_read; inspect each preview/profile before stage=rig with generation_job, creature_id, reviewed_anatomy=true and optional corrected profile. Reuses the anatomy-aware Blender rig, never humanoid autorig. Exports GLB, Blender source and diagnostics. Procedural starting clips are drafts; missing fly/swim/action clips are reported. No automatic game-ready approval. Existing owned immutable source/concept revisions can be reused. Use generation_cancel/retry for durable work.

Plan then build a straight-street district from 1–8 distinct Meshy/Tripo buildings, each manufactured once and reused as separate editable instances. Includes road, pavements, spawn, daylight and camera in the shared Map SceneDocument. Optional owned PBR road/pavement maps; otherwise simple solid surfaces. Plan returns layout, provider-credit count and account quote without spending. Build requires a budget and returns a persistent job, then Map Studio link. Private work version requiring visual review; no automatic Community publication or game replacement.

Persistent concept → isolated handle/segment references → Meshy PBR components → FlexibleChain XPBD → Blender assembly/rig → actual GripForge renderer review. Read flexible_weapon_schema first. recipe stage=concept returns an image before 3D; build can reuse that concept and owned pinned components. Up to 12 existing source clips or sampled motions; character choreography is not generated by this tool. Returns job_id immediately; poll generation_read for weapon GLB, Blender source, physics reports, separate character/weapon review scene and Studio links. Max 23 credits when all art is new; reused components are free. Cancel/retry retains successful provider and fabrication steps. Always private work: no automatic replacement, promotion or Community publication.

Manufacture a NEW fps-arms work asset using the native FPS template: curved hand surface refinement, fitted skinned long sleeves/cuff, optional tactical glove reinforcement and embedded PBR. recipe={version:1,name,quality:high|standard,gloves:tactical|light,sleeves:long|none,colors?:{fabric:#RRGGBB,sleeve:#RRGGBB,trim:#RRGGBB},source?:{assetId,revisionId,fileRole?}}. No source gives standalone hands with native finger rig, hand_item_r/l sockets and existing starting clips. Optional owned immutable native paired source keeps weapon, camera and animation tracks; source must contain FPS_Hands or FPS_Arms_Mesh, all ten native fingers, embedded GLB <=32 MiB. This is template geometry/garment manufacture, not prompt-to-new-anatomy or arbitrary rigging. Returns job_id; poll generation_read. Inspect fingers, seams and weapon/garment contact in the actual renderer before binding. No external provider charge; never auto-promotes or overwrites.

CREATE a game-ready prop GLB (kart, crate, barrel, banana…) and save it to Library as kind=prop. Default Meshy, no T-pose, no biped rig. polycount defaults to 4000. Quoted against the shared subscription, or 10 GripForge credits if it cannot cover the complete operation. Returns a persistent job_id immediately; poll gripforge_generation_read, then download and import the completed model. provider=compute (GripForge Compute, only when asked): stage=quote (free) then stage=generate with quote_token + confirm=true; router type, size, several assets (items), text → four images to validate (gripforge_compute_validate) unless auto_pick; returns studio_url.

Read vehicle_schema first. Recommended recipe.workflow=body_wheels follows concept → references (five wheel-less body views + wheel) → geometry (stop/review) → topology (stop/review) → build (PBR + wheel + rig). Each geometry review uses scene_review with the returned assetRef and must be approved before the next paid step. Rigging preserves UVs and corner normals, GLB/FBX/Blend and measured Chaos setup. Top view is review-only. parts.body.geometry accepts unfinished Hunyuan/Blender GLB; parts.body.topology pins the reviewed reduced GLB; parts.<role>.source reuses finished meshes. Legacy modular_doors remains available for articulated doors. stage=plan quotes each stage without spending; paid stages require its budget. Durable job; poll generation_read. Work draft, real-render review and target-engine driving test required; no automatic publication/replacement.

CREATE a game-ready weapon GLB (knife, gun, sword…) and save it to Library as kind=weapon. Default provider Meshy, no T-pose, no biped rig. polycount defaults to 4000. style=melee|gun is written to meta.style for gripforge_attach. 10 credits. Returns a persistent job_id; poll gripforge_generation_read. Do NOT write one-off Meshy scripts.

Plan then manufacture 1–64 original static weapon models using the existing Meshy PBR pipeline. plan is free and returns an exact aggregate quote and plan_hash. build requires that hash, an explicit budget covering the quote and a stable idempotency_key. One durable job, two bounded fabrication lanes, per-weapon checkpoints and billing; cancellation/retry preserves finished work and provider task IDs. Results remain private work assets. Inspect geometry, textures and actual renderer output before publishing. Does not create attack/reload animations, grip bindings, or flexible-chain physics. Poll gripforge_generation_read; result.items contains Library IDs and Studio links.

Request worker cancellation while retaining completed steps and artifacts. Closing a page alone does not cancel a job.

List workspace jobs. Jobs and completed steps survive closing the studio or restarting the web server and worker.

Free quote for a model, concept, texture, rig or animation. Reports the account funding mode, available balance and estimated operation cost. A subscription combines AI and paid asset generation in one monthly limit without also charging generation credits. Quote every missing role, propose the plan, and obtain a generation budget before submitting paid jobs. Estimates cover one operation, not the full concept/rig/review pipeline.

Poll a job returned by a generation tool. Includes durable steps, progress, failure/retry state and the resulting work revision with a Studio link. include_preview returns saved HUD/map/building concept images (at most 4). Set preview_view to return one chosen view. Concepts are generated illustrations, not verified 3D renders. Success does not mean visual approval or promotion.

Retry a failed or cancelled job, reusing completed steps. An unconfirmed interrupted provider call may be executed again when explicitly retried.

Journey — while building: what the contract still misses, leaves unverified, fails or blocks. The genre defects come from gripforge_game_audit. Read the contract, requirements, implementation status, build-specific evidence, blockers and remediation. Without a contract returns the schema and a proposal. 0 credits.

Local optional Bevy neural snapshot, not a saved scene edit. Requires an open authorized browser runtime session. Configure off/preview; realtime_experimental is refused. Trained weights and a local rights manifest must be selected in the browser; none are downloaded. render_preview starts an asynchronous request: read the command receipt, then query get_render_capabilities for actual completion, model, effective resolution, provenance, timings and limitations. Cancellation prevents stale publication. No images are uploaded. 0 credits.

Body capsules (head/torso/limbs) plus weapons[] — one reach capsule per gf_prop_ wrapper on a loadout/armed bind. Pass id of a loadout GLB (or character_id / attach_id). 0 credits. Use weapons[] not the single weapon field for dual-wield.

Required completion gate: action=workflow stays ready=false until a current runtime capture passes all artistic criteria. After screen_assets/assembly, submit screen_item (or design/project), renderer_version, capture_item, capture and the authorized review_budget; workflow schedules the reviewer automatically. Poll workflow; correct and recapture if changes_requested. No paid retry loop. Genre-independent workflow: pass design (from plan, edited) WITHOUT project to get/quote/concept/review_quote/review. get returns the normalized design hash; quote then concept returns reference_item. For review supply design, reference_item, renderer_version (fingerprint of actual presenter/build including styles/assets/dependencies), capture_item and capture. Forward exact quote.next and poll the generation job. This reviews an immutable snapshot; compare returned identity with current files before reusing the result. No installed kit or per-genre adapter is needed for concept/review. Project mode remains available: Artistic QA: get returns renderer_version and the latest review with current/approved flags. Upload a fresh viewport screenshot to the private Library, then review_quote with capture_item + capture context; submit its exact next payload with review, poll the generation job and read get again. Corrections, stale evidence, missing criteria or unavailable vision never count as approval. Review applies only to the submitted state; preview evidence cannot approve map contrast. v2: action=templates lists reusable compositions and independent themes; plan derives a per-game design from the project art direction and actual capabilities; get reads it; apply writes the same editable document to hud.tactical and ui.theme with expected_revision; quote then concept queues an optional paid visual board. Poll the returned generation job. Studio edits the same document. Live web adapters: FPS hud.tactical, or a shared document presenter for health/mana/XP bars, currencies, quest trackers, panels, text and images using explicit bindings. Other widgets and native document presenters require integration. Never flatten the UI into the concept image or invent gameplay widgets. Omit action for the legacy survivor export. Generate a ready-to-drop survivor HUD: hud.json (layout/anchors, no pixels), original GripForge cartoon PNGs (hearts full/half/empty, portrait frame, currency chips, empty hex slot), a Godot 4 .tscn (Control with named nodes Health/XP/Portrait/Currencies/Minimap), and Unity / Unreal snippets. 1 credit per successful generate. CONSUMER RECIPE — (1) call gripforge_hud; (2) write the PNGs (download the asset URLs) + hud.json (+ hud.tscn) into the game repo, same folder; (3) Godot: instance the tscn, add_child, then get_node("Health").set_health(current, max) and set XP.value; (4) Unreal: the GripForgeHUD plugin must already be in the project (unreal_plugin_url) — point UGFHudWidget::CreateFromFolder at the folder and call BindHealth on your health component. DO NOT redraw the hearts. DO NOT open a UI editor. This hosted endpoint returns file CONTENTS + URLs; pass out_dir only with the npm MCP package, which writes the files for you.

Forge an ORIGINAL themed health-bar style: an image model draws the ornate empty frame (transparent background), the inner trough is located pixel-precisely, and a thematic gradient fill is composed to match — returns an aligned frame+fill PNG pair, same contract as the built-in gripforge_hud bar styles. Describe a THEME (e.g. "gothic demon-hunter, baroque metal, horn ornaments"), never a specific game — styles of a genre, not copies. 2 credits. RECIPE: download frame_url + fill_url next to your hud.json ; Godot: TextureProgressBar (texture_under=frame, texture_progress=fill, custom_minimum_size=size) ; Unity: frame behind, fill cropped horizontally to the health percent. Generation has model variance — if the trough is not found, retry with a clearer theme.

Not the survivor generator (that is gripforge_hud): this one READS an existing HUD. action=import {text, filename?, name?} → reads the HUD of the project and saves it as a HUD document of the workspace; Pass screenshot (a capture of the screen while the game runs) so My HUDs shows the real thing; action shot attaches one to a HUD imported earlier.returns the studio URL to open it. `text` is the engine file as-is: a Godot .tscn scene, a Unity .uxml (UI Toolkit) or a .unity / .prefab (RectTransforms), the Unreal UMG designer clipboard (select the widgets, Ctrl+C — .uasset files are binary), or a GripForge hud.json / hud.v2.json. Read the file yourself and pass its content; `filename` helps detect the format. Controls are recognised by name (health, mana, minimap, ammo, quest…) and type, placed on nine anchors; the report lists what was created and skipped. action=list → HUDs of the workspace. action=get {id} → one HUD document (hud_… id).

Godot InputMap for an FPS: move_forward/back/left/right bound to WASD AND arrow keys, plus jump/crouch/fire/reload/buy. Write res://gripforge/input.json and autoload GfInput (snippet_godot). Player still uses Input.is_action_pressed("move_*"). Optional third_person=true adds portable character-facing JavaScript: movement-facing, independent camera, aim/fire/build facing. Orientation only, not a physics controller. 0 credits.

THE FEEDBACK LOOP: call this after every generate/attach/rig to SEE and VERIFY what you made, before building on top of it. Deterministic, free, repeatable. Returns (1) a structured report — kind guess, triangles/vertices, materials+textures, skeleton (bones, finger bones per hand, cape grid), animations, size in metres — and (2) QA checks with ok/warn/fail + why: plausible scale, poly budget, normals/UVs, rig skinned, strict T-pose arms, five-finger bones, feet on ground. Also returns a multi-angle render sheet (front/right/back/left by default) so you can look at the asset. Fix warns/fails (regenerate, rest_pose, retarget…) then inspect again.

Generate input.joystick source files and installation instructions: touch + mouse, independent dual sticks, radial deadzone, continuous holding and release/focus reset. Unreal runtime Pawn component, Godot CanvasLayer scene, Unity UI component. Feeds the existing movement controller rather than replacing physics, collision or networking. Returns files with relative path/content; the agent writes them into the target project and compiles/tests there. Does not install automatically, modify project settings or claim runtime validation. 0 credits.

A game kit IS an engine: its manifest carries asset slots AND GameRules — the genre as a document (movement, camera, world, objective, entities, pickups, ai, hud, audio). action=list → kits of the workspace. action=get {id} → manifest + rules (lib_… id). action=create {style?, name?, rules?, slots?} → new kit, rules default to the style then merge yours. action=patch {id, rules?, slots?, prompts?, name?} → partial update; rules merge section by section, a section set to null resets it. Slot refs inside rules must point at slot ids of the kit (unknown refs are dropped, the renderer falls back to procedural). Compose a NEW genre by writing rules: e.g. movement vehicle-boost + world arena-ring + objective knockout = sumo cars, no code.

Plan a playable level from a prompt (0 credits): rooms placed in metres (entrance, combat rooms, key/lock, boss room, treasure, secret), connections, critical path, shortcuts, and gameplay spawns (player/packs/boss/rewards). Same seed → same level. Pair with gripforge_scene_kit for dressing props.

Get one Library item (metadata + signed file_url). A forged mesh shows its game-ready verdict first (meta.gameReady: ready, missing checks; meta.collider). VFX items also return studio_url to open the editable effect in GripForge, and workspace_id. A GripForge Compute asset also returns `compute` {edit_url, family: original + colour / material / polygon versions}. Pull bytes into a repo with the npm package.

List items in the GripForge Library locker (characters, enemies, weapons, props, textures, HUD, binds, vfx). Filter by kind and/or game style id (e.g. souls-like, open-world-adventure). Every item carries `rank` {score/100, grade A–D, reasons}: fitness for a GripForge game on the target platform. sort=rank lists the best fit first — use it to choose between candidates. engine filters on the generator: engine=compute lists GripForge Compute assets (manufactures and variants), each with a `compute` block {level scene/detail, family, triangles, edit_url, file_url}. Same origin as Studio — not a marketplace.

Derive a lighter version of a static model (prop, weapon, equipment, vehicle) for a target: decimated to `triangles`, textures bounded to `texture` px, saved as a NEW Library item linked to its master (meta.lod_of). The master is never changed — generate high once, derive mobile/desktop versions. Wheel pivots and nodes are kept. 0 credits.

Return a signed same-origin file URL for a Library item you own (free). This hosted endpoint cannot write into your agent's working tree — use npx @gripforgeai/mcp gripforge_library_pull for that (that is the sync). Community assets are free to clone into your locker unless their creator sets a price in credits.

Upload a file into the Library locker from a URL (or small base64). Local paths need the npm MCP package.

Add a game look (e.g. souls-like, open-world-adventure) to tags so gripforge_style_kit can find the item. Does NOT overwrite grip style melee/gun/shield/staff — that stays on meta.style.

Moon / sun / candle point lights + fog + environment intensity for a game look, plus a Godot snippet. 0 credits. Does not generate meshes. Pair with scene_kit placements (candle tags match candelabra xyz).

Loading overlay for a game look. Returns html/css/js (web) plus engines.godot | unity | unreal. With project_id, action quote → accept → get → apply generates reviewed artwork via a persistent job and binds it to the game; creation_id and quote_token resume it. Title is text — do not rasterize. Overlay alone is 0 credits. Pass character_id (lib_… character/enemy/bind) to generate a cinematic 16:9 background still of that character (1 credit). No letters in the still.

Build a multi-weapon character in ONE call: attaches every prop of every set onto the same rig, names each prop node gf_prop_<slug>_<l|r>, stamps glTF extras { gripforge: { prop, set, hand } } and a scene manifest { default, sets }, then retargets the standard clip pack (animate: false to skip). The engine switches weapons by toggling node visibility. Example sets: { "sword": [{"prop_id":"lib_sword"}], "guns": [{"prop_id":"lib_gun","hand":"left","style":"gun"}, {"prop_id":"lib_gun","hand":"right","style":"gun"}] }. Credits: one attach per prop.

Turn an image into a seamless (tileable) texture: 50% wrap offset, then the seams are cross-faded in a narrow band — the pattern is PRESERVED, nothing is generated or invented. Pass image_url (https) or texture_id (Library id, lib_…). Optional size forces a power of two (512/1024/2048), resampled before the seam fix. The result is saved in your Library as a texture and its download URL is returned. No credit consumed.

gpt-image-2 draws a top-down plan schematic (solid white walls on pure black) + a 3/4 aerial greybox concept render, auto-traces the plan into MapPlan wall boxes and saves both images to the Library. Reuse the traced plan with gripforge_map_plan { item } or https://gripforge.ai/map-plan?item=lib_…&greybox=1. Use when the layout should come from an image instead of the code presets. ~2 images of generation cost.

Queue a persistent map job. stage=concept creates a private master overview plus optional views and stops. views: overview,top_down,entrance,objective. All additional views are conditioned on the SAME pinned master; they are illustrations, not exact 3D renders. reference is an environment screenshot for a NEW layout. To add views to an existing concept, pass source_concept instead; the master is kept unchanged and is free. One credit per new image (full new 4-view set = 4); default is overview only. Poll generation_read with include_preview=true to see the views. After the user selects the concept, submit stage=build with the returned concept and a new idempotency key. Build costs 1 credit and is the default for existing callers; legacy concept_item accepts a Library image ID, while concept pins its exact version. wizard: look stylized|painted|toon|anime|realistic|lowpoly|pixel (rendering, separate from style; omit for the theme default); style realistic|stylized|lowpoly|handpainted (legacy wording); world open_world|closed_arena|dungeon|linear|battle_map; terrain auto|plains|desert_canyon|forest|snow|island|volcanic; size small|medium|large; boundary fixed|infinite (open edge, finite terrain); spawn single|teams; teams 2|4; boss,safeZone,extraction booleans. sources selects catalog/workspace/community props. Build uses the common Scene Engine with independent terrain, regions, paths, zones and asset instances. Cancel/retry keeps successful images and billing. No automatic visual validation, promotion or Unreal import.

Queue a one-time materialization of an owned legacy terrain map into the shared Scene Engine. Preserve the saved spec, seed, resolution, family replacements, placed assets and removed props. Completed jobs return a canonical Studio link; source edits during migration cause a conflict rather than data loss.

Queue a persistent import of schema gripforge.unreal-studio-import.v1. Upload files via local gripforge_map_unreal_upload_local or POST /api/v1/maps/import/unreal/assets first. document is a common SceneDocument with workspace-owned immutable assets, separate props, static foliage patches and editable terrain. source={engine:unreal,level:/Game/...,sha256,warnings:[]}. Returns job_id; poll generation_read for studio_url. Always a NEW private work scene; no overwrite, automatic validation or Community publication. 0 credits; storage quota applies.

Read the native export → binary upload → persistent map import workflow, limits and explicit conversion caveats. Native extraction requires the local MCP and an installed Unreal editor. Hosted MCP/API consume the resulting portable SceneDocument; they cannot read your local .umap/.uasset files.

Reproduce a reference still of the built-in reference map WITHOUT copying pixels. Returns camera (pos/target/fov) + volume/prop/wire ids for that shot. shot=shot21a (A site), shot1a (spawn yard), shot5a (catwalk). Official URL is a LINK only — never rehost, never Meshy image-to-3d. Pair with gripforge_map_plan + map_props + map_wires. 0 credits.

For original competitive 5v5 maps use target=tactical: deterministic modular greybox, real controller/bot paths, standing/crouched sightlines, timings and diagnostic links. Optional art_profile applies reusable PBR harbor facades without changing gameplay collision. Human playtests required. Legacy FPS/bomb map in metres. Interprets the overhead as a 3D compound: hollow rooms, corniche/catwalk, A plat + pit, arches/windows, vaulted tunnels, T/CT, sidewalks. Preview: https://gripforge.ai/map-plan?prompt=bomb%20defusal. Pair gripforge_map_props (crates/cars) + gripforge_map_wires (cables). 0 credits. Not a dungeon — use gripforge_level.

Dressing for an FPS map in metres: crates (A default, B big box, cover boxes), doors (mid, long, B), palms at a spawn, barrels, cars. Matches locker props when tagged. Empty slot → generate that prop. Pair with gripforge_map_plan. 0 credits. Preview https://gripforge.ai/map-plan?prompt=bomb%20defusal

Reconstruct an FPS map: collect official stills (HTTPS LINKS, classified by folder) → MapGraph (callouts/chokes) → footprint greybox (no props) → zone parts → camera match → dress. prompt e.g. "Reconstruct map: desert bomb-defusal map with two sites". stage=collect|graph|greybox|zone|validate|capture|dress|all. zone=LONG_A|A_SITE|…. stage=dress keeps greybox volumes and adds original tileable materials + map_props + map_wires (never the owner's pixels). Preview /map-plan?prompt=bomb%20defusal. A known map stays its owner's IP — internal greybox benchmark, not a commercial clone. Never Meshy a screenshot, never ingest BSP/VPK. 0 credits. Greybox preview https://gripforge.ai/map-plan?prompt=bomb%20defusal&greybox=1

Look up reference stills of a known map the user names (A, B, mid, catwalk, tunnels, spawn). Returns HTTPS URLs of their owner. Reference only — do not rehost, do not send to Meshy image-to-3d. Read materials then generate ORIGINAL locker props. live=true HEAD-checks links. 0 credits. Preview https://gripforge.ai/map-plan?prompt=bomb%20defusal

Read the concept-first map workflow, immutable image reference format, costs and an example. Concepts, map work scenes and external engine delivery are separate steps.

Sagging electrical cables for an FPS map: spans with two anchors (metres) + sag + strands. Not a generated mesh — GfMap tessellates cylinders (role=wire). Typical spans: spawn yard, catwalk, mid, A site. Pair with gripforge_map_plan. 0 credits. Preview https://gripforge.ai/map-plan?prompt=bomb%20defusal

Records match.started. Does not spawn a new process. Mutating.

Read one match record.

Party size, teams, maps, ranked flag, join-by-IP. Write res://gripforge/match.json and autoload GfMatch on top of GfNet. GripForge does not host a matchmaking queue. 0 credits.

Matches recorded for a server (SDK or API).

Marks the match ended. Mutating.

Main menu (host/join/quit), lobby listing connected peers where only the host starts, in-game HUD reading replicated state. Pairs with gripforge_net_kit. 0 credits.

The map of a moba project as a gameplay plan: bounds, team bases and spawns, 1 to 5 lanes (the minions' paths), structure slots (towers, inhibitors, nexus), jungle zones and camp slots, objective zones, the divider (river: walkable; chasm, lava, void: crossed only at bridges), crossings, bush zones, walls, camera limits and symmetry. action "generate" builds a fair map (mirrored through the centre) from lanes, divider, size, team_size, camps_per_jungle, objectives, bushes, walls and returns it with a playability report; apply: true puts it in the project (moba_maps, the match's map and team size, the camera turn). action "validate" checks a map you pass, or the project's: every lane open end to end, spawns, structures and camps reachable, nothing out of bounds, the declared symmetry respected. action "set" stores an edited map document (refused while it has errors, unless force). action "get" returns the current map and its report. New maps default to style {theme: "sanctum", detail: "high", seed: 11, relief: "landscape"}. action "dress" previews a reusable procedural finish for the current map: jungle plateaus, river banks and paved fords, terraced bases, walkable staircases, carved structures, clustered forests, stone guardians and lilies; apply: true saves only the map document, keeping its lane layout and camera configuration; relief updates the shared terrain mesh, ground queries and stair collisions; landscape is the default, terraced keeps base terraces only, flat keeps a level arena. Themes: sanctum, wildwood, ashen; detail: low, medium, high; seed makes the finish repeatable. Bound Library models and textures take precedence. The scenery kit dresses walls and bushes with the bound models. 0 credits.

Journey — a MOBA needs distinct heroes; then gripforge_game_audit for the minions, towers, jungle monsters and the rest of a finished MOBA. Turn champions into the playable roster of a moba project (preset "moba"). Pass the Library ids of 1 to 10 champions that carry an ability pack (the items made by gripforge_abilities_generate: rigged model + basic_attack and ability_q…r clips). Writes the moba_heroes and abilities collections (ids <hero>__ability_q), binds each champion to moba_hero_<n> and the player's champion (player, else the first) to player_character. Bots play the other champions. Returns the heroes with their attack and Q/W/E/R. Then gripforge_game_play_url to play. 0 credits.

Custom animations generated from text with NVIDIA Kimodo on GripForge's GPU, retargeted onto a rigged Library character (or the UAL mannequin): spell casts, signature attacks, emotes, anything the standard clip pack (gripforge_animate) lacks. Kit mode turns a MOBA hero concept's basic attack and Q/W/E/R into clips in one call. Call stage=plan first (free) and show the prompts and price (1 credit per clip); stage=build queues a durable job — poll gripforge_generation_read with job_id (a cold GPU takes a few minutes). The result is a kind=animation Library item, one clip per prompt.

Walkable XZ AABB + circular holes (pillars) for a scene_kit room. JSON the engine clamps against — not a baked .navmesh. 0 credits.

GfNet autoload host/join/spawn with server-side hit validation, gf_round.gd for the round + bomb, MultiplayerSpawner/Synchronizer wiring. 0 credits. Godot only: for a web game built on the Game Kits the network model is the installed sync kit (net.claim = claim, net.authoritative, net.rollback, net.state; room config.net {model, sync, physics} is read from the project): use gripforge_gamekit_install / gripforge_gamekit_configure, then gripforge_docs topics net and web-multiplayer (rooms at /play/online/host/<project_id>?players=N).

Analyze a supplied game performance report, or read the latest instrumented kit capture in your authenticated workspace. Separates frame pacing, measured CPU/GPU work, render counters and stale network snapshots; returns evidence, recommendations and explicit measurement limits. Captures expire after 10 minutes or server restart. Explicit reports are not stored. Does not edit, deploy or automatically fix a game. 0 credits.

HIGH IMPACT. Requires the Unity SDK. Currently returns 501. Owner only.

Read one player session on a server.

Requires the Unity SDK. Currently returns 501. Mutating.

Sessions reported by the Unity SDK. Empty until the SDK calls player.joined.

Durable Blender preparation of an owned static segmented GLB. Map each Door_FL/FR/RL/RR or Wheel_FL/FR/RL/RR to exact mesh node names from the segmentation report. Keeps geometry placement, UVs and PBR; normalizes +Z forward/metres using yaw and length; adds hinge origins, seat/entry sockets and portable GLB open/close/preview clips. Optional hinge coordinates are in normalized vehicle metres. thickness adds an inward door shell, not a full cabin. Optional cuts provide strictly convex door outlines [forwardZ,heightY] and depth [near,far] measured from the centre on that side, in normalized metres. cuts[].meshes restricts each cut to exact exterior/glazing meshes so seats and dashboard cannot be cut. removeMeshes discards explicitly named obsolete trim in the work copy only. Optional clearanceDepthM clears scoped inner walls behind the door skin; trim adds measured body frames, seals and moulded cards below cardTopY. framePoints follows outline in normalized [x,y,z]. Optional trim.fitToSurface projects frames, seals and the inner card onto only the explicitly selected opaque door skin, preserving the supplied outline. attachments adds scoped [min,max] [x,y,z] volumes for mirrors extending beyond the door profile. Blender cuts those explicit profiles while interpolating UVs. Does not infer outlines or guarantee clearance. Returns job_id, new private work GLB, Blender source and animated Studio review link. 0 credits; storage applies. Inspect before replacement.

Read a dedicated-server project (gpj_…) in the authenticated workspace. Not a Library asset.

ADMIN ONLY. Create an editable 3D projectile: ignition, elemental volume, world-space trail, embers, contact flash, oriented ring and soot. 20 named variants: darts, short bolts, orbs, wisps, carrot rocket, spinning disc, spears, comets, coils and lightning. config.variant selects the geometry, trail and impact; style/color optionally override its palette. Omit variant to retain the original elemental recipe. To save several catalogue entries, pass variants=[IDs] without config/name: one persistent job renders/reviews each item sequentially, preserves completed items on retry and returns result.items. Publication remains explicit. Uses the shared trail and Scene Engine modules. save=true (default) queues persistent rendering/review/workspace save; poll gripforge_generation_read for Library ID and Studio link. save=false returns an unsaved recipe. Bind the resulting VFX to combat.projectiles/projectile_vfx using the Game Kit tools; gameplay speed/gravity/damage stay in the module. Web player only in v1. Does not install a kit or modify a game automatically. Recipe compilation needs no AI; saved previews use the existing review pipeline. No automatic visual validation.

Read render.realism systems, modes, bounds and immutable-source workflow for an accessible scene.

Server-side software preview (PNG) of a Library GLB so the agent can see pose/attach without a browser. Not a path-traced studio shot.

Local optional Bevy neural snapshot, not a saved scene edit. Requires an open authorized browser runtime session. Configure off/preview; realtime_experimental is refused. Trained weights and a local rights manifest must be selected in the browser; none are downloaded. render_preview starts an asynchronous request: read the command receipt, then query get_render_capabilities for actual completion, model, effective resolution, provenance, timings and limitations. Cancellation prevents stale publication. No images are uploaded. 0 credits.

Resolve the saved contract, transitive capabilities and rules into an inspectable requirement graph. Does not run tests. 0 credits.

Lower the character's arms from T-pose until they would clip THIS mesh (cape/chest/hips), then save an arms-down rest GLB (clip "idle"). Armed binds keep attached weapons in the hands. 0 credits. Attach first, then rest_pose on the bind id.

Convert animation clips from skeleton A (source_id GLB) onto skeleton B (target_id). Bone-name heuristic (Mixamo / UAL / Tripo). Saves kind=animation.

Close and delete one online room (frees the room quota). Invitations to it stop working. Mutating.

List the online (multiplayer) rooms of the workspace with their live state. A closed page can leave a room open until its lifetime ends; each workspace holds at most 8 live rooms, so list and close the stale ones when "Room capacity reached" appears.

Queue a versioned GripForge runtime command for the exact browser session. Use inspect, apply_patch, set_relief (off/pom/spom on configured relief surfaces), step (paused only), run_test, play, pause, reset, load_game, unload_game or get_stats. Returns received; fetch the applied receipt with gripforge_runtime_receipt. Preview changes do not persist the project. Unknown kits/components and invalid fields are refused by the Rust core. 0 credits.

Read received/applied/rejected status for a Rust preview request, including simulation boundary and revisions. 0 credits.

List live Rust/Bevy previews for an authorized Game Kit project. Open the project Rust player in GripForge first. No server-side solo simulation. 0 credits.

Archive the owned Library resource bytes and metadata for use in scenes. Returns assetId/revisionId, capabilities, material/animation names and sockets. Repeating for identical content reuses the archive; it consumes workspace storage only once. Prefers the validated version for insertion; choose version=work to test the draft. Does not approve the asset visually.

Promote only when the exact work revision has a matching trusted review for the current renderer. Existing scene instances keep their pinned revision.

Create a shared test scene and queue an independent review of real GripForge captures. Supports meshes, textures and effects. Never promotes automatically.

Read immutable asset work/current pointers and the trusted review of work. Does not create a version.

Create a draft scene. Pin Library assets first and use returned assetId/revisionId references. The response contains a GripForge Studio link. Exporting a VFX definition must not include this test scene.

Return locker props for a game look plus a suggested layout (metres, Y-up). Call after style_kit. Empty slots mean generate that dressing prop, tag it, retry. Does not generate meshes.

List scenes, work/validated revision pointers and links in the authenticated workspace.

Queue a scoped edit of saved scene nodes or a region. The server derives the allowed scope and preserves locks, gameplay zones, outside geometry and the last validated revision. Blender may fabricate missing assets. The result is a proposal requiring explicit application, never a silently replaced map.

Queue resumable browser texture variants for a saved scene. Preserve original GLBs, geometry, transforms, PBR, animations and Library heads. Color textures up to 1024px, data maps up to 512px and terrain color up to 4096px; WebP encoding. Saves a new work revision only if expectedRevision still matches. Does not promote validated. Poll generation_read for byte savings and studio_url. 0 credits; storage quota applies. New Unreal imports do this automatically.

Apply insert/remove/transform/replace/update commands atomically. Environment updates can configure shared 3D cloud layers (density, altitude, wind, colour and quality); see gripforge_scene_schema. Replace preserves id, transform, role, materials and attachments, and rejects incompatible named slots. Locked objects cannot be changed. Scope constrains regional edits; read the schema first.

Explicitly promote the latest work revision only if its trusted visual review matches the exact content, dependencies and current renderer. expectedCurrent must match the existing validated pointer or null. No client-supplied approval report is accepted.

Explicitly accept or discard a proposal after inspecting it. Accept requires expectedRevision to match the exact base and current work revision. Atomic: concurrent changes or discard abort the save. Creates work, never promotes validated.

Read the original scene, proposed document and scoped commands. Use the Studio link to inspect actual rendering before applying.

List persistent proposals and their base revision, status and descriptions.

Read the editable SceneDocument at work, validated, or a numbered revision. Keep work_revision for optimistic writes. Validated can be absent.

Apply reusable architectural surface profiles and GTAO to a saved work scene. Read gripforge_scene_schema.realism first. assignments [{nodeId,material:"*" or a named slot,surface:{preset:preserve|brick|stone|metal|glass,detail:0..1,weathering:0..1,regions:[]}}]; surface:null removes only this local profile. Optional glass regions use explicit normalized mesh-local min/max bounds and interior parallax settings. Never guess masks from colours. ambientOcclusion {enabled,radius:0.02..3 metres,intensity:0..1,quality:low|medium|high}, or null to remove. Preserves transforms, source PBR maps, other instances and validated revisions. Returns Studio link for visual comparison. Optional quality {preset:performance|balanced|quality,resolutionScale:0.5..1,sharpness:0..0.5}, or null to restore the original presentation. Quality supports viewport resolution up to a bounded 4K budget. Optional automotive {look:studio|daylight,quality,lengthM,depthOfField:false,focusDistanceM} explicitly applies a whole-scene look: HDR strip reflections, contact AO, quality and restrained post-processing. Studio replaces an existing HDRI; daylight keeps it. Optional colorGrading/depthOfField overrides or null. DOF is perspective/capture only, disabled with volumetric passes. No generation charge, DLSS or automatic approval.

Queue a persistent render and independent visual review of the exact saved work revision. Poll generation_read. Optional assetRef records the review on the exact isolated asset revision as well; required for vehicle geometry/topology gates. This does not promote or replace the validated version.

Start here. Read the shared scene model, kinds, commands, limits and an editable example. Assets are immutable revisions; placed objects are independent instances. Scene saves create work revisions, never automatic visual approval.

Read work/current pointers, immutable revision history and trusted visual review evidence. A rejected or unavailable review does not change the validated version.

Save a complete SceneDocument. expectedRevision must equal the latest work revision. A 409 is a real conflict: reread before changing it. Does not change the validated revision or source Library assets.

Create a dedicated server record. Godot supports ENet and gfnet; gfnet requires a game server Docker image that boots its simulation and validation hooks. Does not start the process until deploy.

HIGH IMPACT. Stops the instance and deletes the server. Requires confirm=true. Owner only.

Create a Docker instance for this server and start it. Local demo uses a real container (python heartbeat unless a custom image is set). Mutating.

Activity feed: server.created, server.ready, player.joined, match.started, …

Read one dedicated server, its instance, engine/networking detection and integration status.

List dedicated game servers in the workspace with current status and latest instance.

Stdout/stderr from the runtime provider (Docker logs). Not stored in SQL.

CPU, memory, network counters and player count from the metrics provider.

Restart the running instance. Mutating.

Update desired max players. Does not resize cloud hardware (Docker provider has no node pool). Mutating.

Start the existing instance. Mutating.

Stop the running instance. Mutating.

The workspace IS the game: genre (a game kit style id, e.g. arena-car), remappable controls (same InputKit as the Godot input.json AND the web players), match rules (rounds/laps/score/timers/ports) and menu options (title, HUD toggles). Once set, every *-kit route called WITHOUT ?prompt= reads them (explicit prompt always wins), and the web players use the same bindings. Call with no args to READ (settings + resolved defaults). Pass fields to MERGE-SET. Pass a section as null to reset it to the genre preset. Set genre at project start, right after gripforge_style.

Pull a catalog shader (id or alias like slash_reveal) for godot | unity. Three.js is no longer a supported target. Returns file contents + assignment snippet (Godot ShaderMaterial / Unity Material). This hosted endpoint cannot write into the working tree — use npx @gripforgeai/mcp gripforge_shader_pull with out_dir for that. v1 pull is free (no credit).

List shaders from the GripForge catalog (authorized GodotShaders.com mirror). Returns { id, name, engines }. Search q=slash_reveal. Then pull with gripforge_shader_pull. Free.

ADMIN ONLY. Compile a shield recipe into reusable native source files for Godot 4, Unity or Unreal 5, using the same geometry and shader functions. Returns files keyed by relative path plus install instructions. Write ALL files into a dedicated directory; never overwrite an existing modified package without checking. Includes enable/hit/shatter runtime; no mannequin or preview environment. Native builds require target-engine verification; source export is not visual certification.

ADMIN ONLY. Create an original editable 3D elemental shield using the shared shield-vfx module: 17 styles, radius/stretch, colours, intensity, quality and seed. Same shaders used in Studio and native exports. save=true by default queues persistent rendering/review/workspace save; poll generation_read for Studio URL. save=false returns an unsaved project. No commercial reference assets, collision or damage. Recipe compilation needs no AI; saving uses the normal VFX review pipeline. Work is not visually validated automatically.

The visual identity of the workspace: palette (hex colors), written art direction, and sliders (realism/saturation/edges). Once set, it is AUTOMATICALLY appended to every generation prompt (characters, enemies, props, HUD bars…) so ten assets come out as ONE game — and gripforge_inspect gains a style check flagging assets whose dominant colors drift from the palette. Call with no args to READ the current DNA (+ the compiled prompt block). Pass fields to SET it. Pass extract_from (image lib ids) to auto-build the palette from reference art. clear:true removes it. Set the DNA at the START of a game project, before generating.

Resolve a game look the user names ("stylish gothic action", "anime open-world adventure") to a game-style id, then return the locker kit already tagged with that look: characters, enemies, weapons, props. Call this BEFORE generating when the user wants that game. Reuse the returned Library ids (attach / pull). Only generate if a bucket is empty.

Make a map of the terrain studio the world of a game that has the world.terrain kit. Pass map (a terrain studio map id, lvl_…) or scene (a scene id, with terrain when the scene has several terrain nodes). The map is baked once into a file (heights, ground roles, lakes and rivers, vegetation, roads, spawn and objective points) stored in the Library and bound to the terrain_map slot; the project data keeps its reference in terrain_maps. The same file is delivered to Godot, Unity and Unreal with the kit. Buildings and other non-vegetation props, and terrain holes, are not carried yet (counted in map.notCarried). Without a bound map world.terrain keeps generating its world from the seed. 0 credits.

Journey — after each change, with gripforge_game_test (the headless playtest). Passing tests are not a finished game: gripforge_game_audit, then gripforge_validate_release. Queue validation on the existing durable job worker. Plays the game with its own controls: the suites each kit declares (several clients joining, leaving and coming back, click orders on the ground, an enemy, a door, a chest, a downed ally, held buttons, a player that plays to an ending) and the headless playtest; endings (victory, defeat, boss defeated) come back as events. A real page that reached an ending, its hardware frame rate and the default screens it opened count for this revision. Returns job and status_url; poll that URL with the existing job tools. Simulation is not end-to-end gameplay proof. 0 credits.

Prep a game texture (terrain / props): optional 50% wrap + seam blend, then a faithful lanczos upscale (1× / 2× / 4×, capped at 1024 or 2048). The pattern is PRESERVED — nothing is generated (not Magnific, not an image model). Pass texture_id (Library) or image_url (https). Result is saved as kind=texture and the albedo URL is returned. To write the PNG into an open repo, use the npm package with out_dir. No credit consumed.

Reusable death/victory screen concepts backed by the existing ui.endscreen Game Kit. List four visual templates, validate or export screen.json + a standalone web/Bevy presenter + runnable preview. No credits, no AI image generation. Connect authoritative game events and acknowledge multiplayer retries; a downed player is not a party defeat. Get health PNGs from gripforge_hud.

Free, deterministic parametric architecture. plan validates a metric recipe; build queues a durable common-compute job and returns job_id to poll with gripforge_generation_read. Real window reveals, balconies, shopfronts, rooftop HVAC, photographic PBR and optional arid shutters/canopies/water tanks. Private immutable GLB plus recipe/report; reuse via Library bindings or scene instances. Nominal width/depth exclude protrusions: use returned actual bounds and preserve proportions. A solid envelope with storefront depth, not traversable interiors, a complete city or an AI replacement for gripforge_generate_building. Review in the actual game. 0 credits.

Journey — last, after gripforge_game_audit says complete. The game is finished only when report.ready=true on a contract that holds what the user chose for the game (genre proposals taken included); otherwise tell the user what is left and why. Evaluate explicit release blockers for the current build. Only report.ready=true proves all mandatory acceptance checks passed; coverage ratios never override blockers. 0 credits.

Save an independent door pose on one prepared vehicle instance in a saved scene. doors maps Door_FL/FR/RL/RR to 0 (closed)..1 (open). Unspecified doors retain their state; null clears the manual pose. Requires exact expectedRevision and named hinges/portable clips from prepare_vehicle. Pauses the global animation for manual control. Preserves the source asset, transforms, other instances and validated versions. Returns the actual Studio link. No generation charge.

Repair generated automotive surfaces. Preferred: source (immutable static GLB with Body/Wheel_* groups) + repair recipe queues a persistent Blender job; guided glazing/panel regions, paint cleanup, coloured caliper extraction, measured replacement panels/curves with PBR and fitted boundaries, optional rebuilt wheels and rigid wheel rig. features.marks adds surface-fitted pinned PNG logos, relief badges and mesh lettering; wheel markings follow their wheel, named authored marks are replaced in the new work copy. For grilles or trim on a reconstructed recess, set curves.fitSurface.panel to its non-structural panel name to follow the final part. Returns a NEW private work version with GLB/Blend/optional FBX and GripForge review link. Does not promote or overwrite the source. Regions use +Y up/+Z forward metres, bounded displacement; region-only glass stays opaque, authored transmissive panes need an authored cabin. Poll/cancel/retry generation-jobs. 0 provider credits. Legacy id without repair uses one-time material/wheel finish and rewrites its Library file.

Yaw so a kart/vehicle GLB nose is local +Z (Meshy often faces +X). Then heading = yawToward(track tangent) = atan2(fx, fz). Pass Library id of the kart prop. Returns meshYaw + Godot snippets. 0 credits.

Read the concept-first vehicle recipe, canonical parts, hinge convention, limits, quotes and revision workflow.

Quote then queue one durable Tripo v2 semantic segmentation of an owned immutable GLB, including Meshy or Blender sources. stage=plan is free; stage=build requires quoted budget. 40 Tripo provider credits, distinct from the GripForge quote. Returns job_id; inspect parts and texture preservation in generation_read. No automatic completion, retexture, articulation, publication or replacement. A segmented output still needs hinge preparation.

ADMIN ONLY. Generate + export a 3D slash/thrust/spin VFX parented to an existing attach bind, or a world-space fire meteor (head + ribbon trail + impact). Pass prompt to fill type/style/color/emitters. Returns studio_url (open this link to edit the saved effect in GripForge), workspace_id, GLB, PNG and vfx.json. Refused for non-admin API keys.

ADMIN ONLY. Text-to-VFX or image-to-VFX: interpret a prompt or reference into an editable animated effect specification. No Library save or credit charge. Pass the resulting type, style, look, color, duration, width, name, emitters and generation to gripforge_vfx to save/export (1 credit). Image mode requires a configured vision provider.

ADMIN ONLY. Play/scrub pose of a strike VFX at time t (seconds). t=0 is empty. No credits. Arc reveal / trail length / burst particles. Refused for non-admin keys.

ADMIN ONLY. Text/image -> durable generation job -> original geometry/shaders -> actual GripForge rendering -> visual critique/correction -> workspace work revision. Returns immediately with job_id, status_url and studio_url. Poll gripforge_generation_read, then open its source_url to inspect the saved source and review. Closing the request does not cancel the job. Completed steps survive worker restart. Pass id + expected_revision to refine; image_id must belong to this workspace. A visually rejected candidate never replaces the validated version.

ADMIN ONLY. Explicitly promote work after a trusted positive review matching this source, dependencies and renderer. expectedCurrent is the existing validated revision, or null.

ADMIN ONLY. Read a project in the authenticated workspace. Returns project.json, per-layer .vert/.frag source, current revision, recent revision history and a Studio link. Read an older revision to restore it through write with the CURRENT expected_revision. Does not grant access to the server repository or arbitrary user files.

ADMIN ONLY. Queue a durable render of an owned id OR an unsaved project at 1–4 timestamps. Returns job_id immediately. Poll gripforge_generation_read with include_preview for the actual rendered image, coverage, clipping and shader diagnostics. Uses the shared GripForge SceneViewport. Saves preview evidence only, no project revision. Requires workspace write access because the job and preview consume storage. Render timings are not browser FPS.

ADMIN ONLY. Queue a fresh GripForge capture and independent visual review without writing another source revision. Poll gripforge_generation_read; inspect its preview, then explicitly promote if approved.

ADMIN ONLY. Start here to create original VFX beyond presets. Returns the editable source format, shader helpers, geometry budgets and a working example. Workflow: write custom source -> render -> inspect the returned image -> revise -> save. Blender or another modeling tool may supply custom mesh vertices/UVs/indices; no Blender instance is required for procedural effects.

ADMIN ONLY. Create with project, or update id with project OR edits:[{path,content}]. Updates require expected_revision. Returns HTTP 202 job_id and studio_url immediately. Poll gripforge_generation_read; the worker compiles, renders and saves a work revision, then reviews it visually. Completed result includes source_url, preview_url and review. Use generation_read with include_preview to see the actual image. A source conflict fails the job and preserves current work; reread before writing. Validated versions remain unchanged. Editable files are project.json and layers/{id}.vert/.frag.

CS-style first-person arms + gun under Camera3D/Hold. Returns viewmodel.json, gf_viewmodel.gd, clip shader. Pull Mixamo T-pose characters (palms down) into bodies paths. Gun is a sibling under Hold, not the world-body hand bone. Pair with gripforge_input_kit. 0 credits.

Reusable voice identities in Community > Audio > Voices and the workspace. list/get return provider, model, voice, tested languages, acting instructions and timed sample ranges in file_url. schema describes the contract. characters reads speakers and revision from an OWNED Game Kit project. To assign, first clone a community voice into the workspace, then supply voice_id, project_id, speaker, language and expected_revision from characters. Optional instructions describe the character acting style. Saves a per-speaker/language snapshot in dialogue.voice voice_cast, preserving other voices and dialogue content. Assignment performs no TTS: bake recordings with tools/dialogue-voice/build.mjs and a matching worker. Per-line emotions use dialogue_voice_directions. Free reads; no GPU rental or automatic generation.

Queue a persistent weapon-aware motion draft from an owned armed GLB. Axe/hammer get distinct procedural cleave, backhand and overhead attacks; sword uses the explicit UAL sword pack. Staff attacks fail with compatible_motion_missing; no silent sword fallback. Holds Grip finger poses, supports a second-hand anchor with IK. Poll gripforge_generation_read, cancel/retry with generation tools. Returns a NEW work asset and Character Studio URL; never replaces the game/source or visually validates a result. 0 generation credits.

Independently sample the saved weapon-motion GLB at 60 Hz: grip drift, finger pose, wrist rotation, joint discontinuities, second-hand reach and torso overlaps. Reports technicalPassed separately from visualStatus=not_reviewed. Inspect idle, wind-up, impact, recovery and locomotion in GripForge before using the draft. No promotion or mutation.

Measure an owned armed GLB instance and return its editable weapon/motion profile: explicit family, holding hand, dimensions, grip anchors, handle and cutting-edge axes. Use gripforge_attach first. Default axes are +Y handle and +Z edge; inspect/correct them. Does not mutate the Library source or certify the initial seating.

Not implemented. Returns 501. Mutating.

Not implemented. Returns 501.

Worlds stored for the workspace or a server’s project. Empty until worlds are implemented.

HIGH IMPACT. Not implemented. Returns 501.

Return the skeleton frame for a limb. Outward = R_Thigh−L_Thigh (never ask the LLM). Up = Knee→Hip / Ankle→Knee. Match getWearableFrame axes — no Euler.

Analyze an armor asset once: up, bodyFacing, outward, anchor. Never returns rotate X = 180°.