Generate faceless content video prompts for Seedance 2.0 on Higgsfield. Use for faceless YouTube channels, TikTok content without showing face, anonymous creator content, narration-driven videos, stock-footage-style AI content, or any video where the creator doesn't appear on camera. Triggers on faceless, no face, anonymous, narration, stock footage, b-roll, background video, voiceover video.
npx skills add https://github.com/rediumvex/ai-video-generator-claude --skill seedance-faceless-channel
When generating a prompt, collect or infer:
| Field | Description | Example |
|---|---|---|
| Topic | Subject matter of the video | "The collapse of Silicon Valley Bank" |
| Platform | Target platform and aspect ratio | YouTube 16:9 / TikTok 9:16 / Reels 9:16 |
| Duration | Clip length in seconds | 5s / 8s / 10s |
| Mood | Emotional tone | cinematic / moody / energetic / calm |
| Style | Visual language | documentary / motion graphics / aesthetic / abstract |
| Sound priority | Narration-first or music-first | narration-first |
| Hook type | Opening pattern | visual metaphor / data reveal / satisfying loop / impossible scene |
Seedance 2.0 prompt structure:
[SCENE DESCRIPTION] — [CAMERA MOVEMENT] — [LIGHTING] — [MOOD/ATMOSPHERE] — [MOTION DETAILS] — [SOUND DIRECTION]
Each prompt block should be 15–25 lines covering: scene, camera, lighting, timing breakdown, motion notes, and sound direction.
The first 2 seconds determine whether the viewer swipes. Use one of these four proven patterns:
| Pattern | Description | Visual Execution | Best For |
|---|---|---|---|
| The Visual Metaphor | Open on a powerful image that symbolizes the entire video's thesis | Extreme close-up of a burning dollar bill, then pull back to reveal a city skyline | Finance, business, history channels |
| The Data Reveal | Numbers or statistics materialize on screen in a visually dramatic way | Counter spinning rapidly then freezing on a shocking number, overlaid on relevant environment | News, education, statistics-driven content |
| The Satisfying Loop | A seamless, hypnotic motion that locks the eye before the brain can decide to swipe | Fluid pouring, sand falling, clouds rolling — perfectly timed to feel infinite | Meditation, ASMR, aesthetic channels |
| The Impossible Scene | A physically impossible or hyperreal visual that creates instant curiosity | A city at the bottom of the ocean, a forest growing in fast-forward inside a skyscraper atrium | Sci-fi, speculative, "what if" channels |
Best for: Documentary, history, finance, crime, geopolitics
Best for: Tech explainers, startup breakdowns, tutorials
Best for: Mindset, productivity, study-with-me, lo-fi, wellness
Best for: True crime, history, investigative journalism, case studies
The voiceover carries all information. Visuals illustrate and punctuate.
The music is the backbone. Narration (if any) drops in between musical phrases.
Topic: The 2008 Financial Crisis
Style: Cinematic B-Roll + Documentary Reconstruction
Hook: The Visual Metaphor
Sound: Narration-first
OPENING FRAME (0:00–0:02):
Extreme close-up of a single red chess king piece tipping in slow motion, falling onto a reflective black surface, camera follows it down in real-time — shallow depth of field, everything behind it blurred into warm bokeh circles — no sound yet except the faint, growing hum of a low string note.
PULL-BACK REVEAL (0:02–0:05):
Camera pulls back slowly to reveal the chess king is surrounded by dozens of other fallen pieces on an abandoned boardroom table — overhead fluorescent lights flicker once — the color grade desaturates to near-monochrome with a faint green tint, as if seen on CCTV.
ENVIRONMENT ESTABLISH (0:05–0:08):
Slow horizontal drift left along the length of the boardroom table — empty coffee cups, scattered papers, a disconnected phone receiver — camera settles on floor-to-ceiling windows revealing an empty city street below at dawn, golden light creeping in through smoke-hazed air.
MOTION NOTE:
All camera movement under 0.3% per frame — the scene should feel frozen in time, held in suspension. No handheld drift. Parallax only between table edge and the city window.
LIGHTING:
Practical fluorescent flicker inside (tungsten, 3200K, slightly green), counterbalanced by cool dawn light (7000K) from window — the two temperatures collide at the table edge, creating a visible color split line.
SOUND DIRECTION (narration-first):
Narration enters at 0:03 over the pull-back. Music bed: low cello drone with a single sustained piano note that decays slowly. No drums. No melody. Ambient SFX: fluorescent buzz, distant city traffic muffled through glass, the faint sound of paper shuffling once.
Topic: How AI is replacing white-collar jobs
Style: Motion Graphics Hybrid
Hook: The Data Reveal
Sound: Music-first
HOOK FRAME (0:00–0:02):
Black screen — then a single white number "0" materializes in the center in 3D, extruded type, floating in dark space — it rapidly counter-spins upward through 10, 100, 1,000 — freezes at "4,000,000" in electric blue, the number glowing and pulsing once — camera begins a slow orbit around the number as particle streams flow out from its edges.
ENVIRONMENT REVEAL (0:02–0:05):
Camera orbits backward and the "4,000,000" integrates into a 3D data visualization — a cityscape of bar charts and flowing data streams — below the number, text materializes in white: "jobs automated in 12 months" — camera continues orbiting at 0.5 degrees per frame.
CONCEPT VISUALIZE (0:05–0:08):
Camera descends into the data cityscape — buildings made of stacked icons representing job categories (lawyer, accountant, coder, writer) — as the camera moves through each district, icons dissolve into particle dust that flows upward, replaced by glowing AI nodes forming connection meshes.
MOTION NOTE:
All 3D elements rotate continuously at different speeds — data streams flow downward in background, job-icon buildings pulse before dissolving, AI nodes expand outward in rings. No static element exists in any frame.
LIGHTING:
All internal — neon blue and violet from data streams, hot white from number display, deep black background. No ambient fill. Hard edge shadows from every element. Color pops: electric blue (AI elements), desaturated grey (human-job icons before dissolution).
SOUND DIRECTION (music-first):
Music: dark electronic with sharp hi-hat pattern — cut lands on downbeat. Counter spin SFX on number reveal (0:00–0:02): mechanical ticker tape sound. Camera orbit has subtle whoosh. Each dissolving job icon gets a brief soft chime. Narration (if used): drops at 0:05 in a musical gap.
Topic: Why waking up at 5am changes your brain
Style: Ambient Aesthetic
Hook: The Satisfying Loop
Sound: Music-first (ambient)
OPENING LOOP (0:00–0:03):
Extreme macro close-up of coffee steam rising from a mug — perfectly backlit by a single warm desk lamp (2700K) against a dark, blurred background — steam curls upward in a natural looping pattern — camera imperceptibly breathes (0.2% push-in over 3 seconds) — the shot feels infinite, like it could repeat forever — no cuts.
ENVIRONMENT DRIFT (0:03–0:06):
Slow horizontal drift right from the coffee mug across a minimal desk surface — a journal open to a blank page, a single pen resting diagonally, a window at the far edge where pre-dawn deep blue light begins — camera arrives at the window and holds for 1.5 seconds.
DAWN REVEAL (0:06–0:09):
Very slow tilt upward through the window — reveal the first thin line of orange at the horizon beneath a navy blue sky — stars still faintly visible — the camera holds this composition — the light level rises almost imperceptibly during the 3-second hold.
RETURN TO DETAIL (0:09–0:12):
Camera drifts back down to street level view through the window — empty street, single streetlight reflecting in a small puddle — focus pull from the distant horizon to the puddle reflection — the warm orange of the sunrise is now captured in the puddle, concentrated and bright.
MOTION NOTE:
Every movement must be smooth enough to feel like a slow zoom on a locked-off camera. Target: a viewer who screenshots this for a wallpaper. No movement faster than 0.3% per frame. No cuts within the clip — all transitions are camera drifts or focus pulls.
LIGHTING:
Entirely practical and environmental. Desk lamp at 2700K (orange-warm). Pre-dawn sky at 7500K (deep blue). The entire clip lives in the contrast between these two temperatures — warm interior, cool exterior — meeting at the window frame.
SOUND DIRECTION (music-first):
Music: lo-fi piano, slow tempo, no drums, prominent vinyl crackle. Ambient SFX layer: coffee steam sizzle (barely audible, 5% volume), distant birds beginning at 0:06, wind trace from 0:09. No narration. No voice. The silence should feel like 5am — the world not yet awake.
TOPIC: [subject]
PLATFORM: [TikTok 9:16 / YouTube 16:9 / Reels 9:16]
DURATION: [4s / 8s / 10s]
STYLE: [cinematic b-roll / motion graphics / ambient / documentary]
HOOK (0-2s):
[Hook pattern] — [opening frame description]
Camera: [movement + speed]
Sound: [impact layer]
BEAT 1 (2-Xs):
[Scene description]
Camera: [movement]
Lighting: [setup]
Sound: [layer]
BEAT 2 (X-Ys):
[Scene description]
Camera: [movement]
Lighting: [shift if any]
Sound: [layer]
CLOSE (final 1-2s):
[Resolution frame]
Camera: [settling movement]
Sound: [resolution]
MATERIAL REFS: @image1 [what], @video1 [what], @audio1 [what]
Follow these constraints every time you generate a Seedance 2.0 faceless prompt:
Hard rules:
Quality rules:
10. Never use the word "beautiful" or "stunning" in a prompt — describe the specific visual quality that creates that impression instead.
Platform-specific:
Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.
Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.
Improves the quality of images, especially screenshots, by enhancing resolution, sharpness, and clarity. Perfect for preparing images for presentations, documentation, or social media posts.
Downloads videos from YouTube and other platforms for offline viewing, editing, or archival. Handles various formats and quality options.
Lightweight WSI tile extraction and preprocessing. Use for basic slide processing tissue detection, tile extraction, stain normalization for H&E images. Best for simple pipelines, dataset preparation, quick tile-based analysis. For advanced spatial proteomics, multiplexed imaging, or deep learning pipelines use pathml.
Microscopy data management platform. Access images via Python, retrieve datasets, analyze pixels, manage ROIs/annotations, batch processing, for high-content screening and microscopy workflows.
Python library for working with DICOM (Digital Imaging and Communications in Medicine) files. Use this skill when reading, writing, or modifying medical imaging data in DICOM format, extracting pixel data from medical images (CT, MRI, X-ray, ultrasound), anonymizing DICOM files, working with DICOM metadata and tags, converting DICOM images to other formats, handling compressed DICOM data, or processing medical imaging datasets. Applies to tasks involving medical image analysis, PACS systems, radiology workflows, and healthcare imaging applications.
This skill should be used when working with pre-trained transformer models for natural language processing, computer vision, audio, or multimodal tasks. Use for text generation, classification, question answering, translation, summarization, image classification, object detection, speech recognition, and fine-tuning models on custom datasets.
Take rediumvex/seedance-faceless-channel from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.