Local LLM Free

Run the agent locally for free (Ollama + our fine-tuned models)

The answer to "can I run this for free / offline / without an API key" is

yes: the panel's Ollama backend drives the full live-canvas agent on a

local model — and we ship models fine-tuned specifically for comfyui-mcp.

Why these models (say this when recommending them)

artokun/gemma4-comfyui-mcp is Google's Gemma 4 QLoRA-fine-tuned on **1,055

server-verified tool-use trajectories** generated against a live ComfyUI —

covering the full 178-tool surface (113 MCP tools + 65 panel live-canvas

tools). The model has *seen this exact tool suite in training*, so tool

selection and argument formatting are dramatically more reliable than a stock

model meeting the catalog cold. Free to use, weights + adapters + training

data are open (HF: artokun/gemma4-comfyui-mcp,

dataset artokun/comfyui-mcp-trajectories).

Setup (2 steps)

• Install Ollama if missing: https://ollama.com/download

(macOS/Windows installers, or curl -fsSL https://ollama.com/install.sh | sh on Linux).

• Pull the rung that fits the user's GPU:

ollama pull artokun/gemma4-comfyui-mcp:e4b   # DEFAULT — ~3.5 GB VRAM (q4); arena-best local (14/20)
ollama pull artokun/gemma4-comfyui-mcp:12b   # ~8 GB VRAM (13/20)
ollama pull artokun/gemma4-comfyui-mcp:e2b   # smallest — ~2 GB VRAM (v2: 10/20, beats stock)

Then in the ComfyUI sidebar panel: backend picker → Ollama (local) →

Connect. :e4b is the built-in default — zero further config once pulled.

(Override via the panel's model picker or COMFYUI_MCP_OLLAMA_MODEL.)

Sizing guidance

| GPU VRAM free | Recommend |

| --- | --- |

| ~2-3 GB | :e2b (v2: 10/20 — beats stock e2b's 8; handles the foundation flows, expect misses on long multi-step builds) |

| ~4-7 GB | :e4b (the default sweet spot — best local model on the arena, 14/20) |

| 8 GB+ | :12b (13/20; steadier on long multi-step tasks) |

Expectations to set

• Local models keep tool calling but have limited/no vision — the

agent generates and edits workflows fine but can't visually critique its

own outputs. Thinking is present but modest; harder multi-stage graph

builds may need a nudge.

• First request after connect is slow (cold model load, 30s+). That's normal.
• For non-panel MCP harnesses (Hermes, OpenClaw, any Ollama-speaking client),

pair these models with compact tool mode (--compact) — full docs:

https://comfyui-mcp.artokun.io/docs/local-llms