Skip to main content
Glama
Sign In
Comfy Pilot - OpenCode Edition

Comfy Pilot - OpenCode Edition

by jaijia
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

Comfy Pilot - OpenCode Edition

MCP server that lets OpenCode (or any MCP-compatible AI coding tool) directly control ComfyUI for AI image generation.

Forked from ConstantineB6/comfy-pilot and adapted for OpenCode compatibility.

What this does

Instead of manually clicking nodes in ComfyUI, you just tell your AI agent:

  • "Generate a 1920x1080 background of a dystopian room"

  • "Create a character sprite with 3 expression variants"

  • "Download this model and set up a workflow for it"

The agent uses MCP tools to manipulate your ComfyUI workflow, set prompts, and run generations — all programmatically.

Related MCP server: ComfyUI MCP

Features

Tool

Description

get_workflow

Get the full current workflow

summarize_workflow

Human-readable workflow summary

get_node_types

Search available node types

get_node_info

Get detailed info about a node

get_status

Queue status and system info

run

Run the workflow / interrupt

edit_graph

Batch create, delete, move, connect nodes

view_image

View outputs from image nodes

search_custom_nodes

Search ComfyUI Manager registry

install_custom_node

Install a custom node

uninstall_custom_node

Uninstall a custom node

update_custom_node

Update a custom node

download_model

Download models from HF/CivitAI/URLs

Quick Start

1. Install dependencies

pip install mcp httpx

2. Start ComfyUI

Make sure ComfyUI is running on http://localhost:8188.

3. Configure OpenCode

Add to ~/.config/opencode/opencode.jsonc:

{
  "mcp": {
    "comfyui": {
      "type": "local",
      "command": ["python", "C:\\path\\to\\mcp_server.py"],
      "enabled": true
    }
  }
}

Or for Claude Code (~/.claude/settings.json):

{
  "mcpServers": {
    "comfyui": {
      "command": "python",
      "args": ["/path/to/mcp_server.py"]
    }
  }
}

4. Restart your AI tool

Restart OpenCode or Claude Code to pick up the new MCP server.

5. Start generating

Once connected, you can just talk naturally:

"Load the Counterfeit model, set prompt to a dark sci-fi room, 1920x1080, and run it."

Requirements

  • Python 3.8+

  • ComfyUI (running on localhost:8188)

  • An MCP-compatible client (OpenCode, Claude Code, etc.)

How it works

┌──────────────┐       stdio/MCP        ┌──────────────┐       HTTP REST       ┌──────────┐
│  OpenCode    │ ◄──────────────────────►│  MCP Server  │ ◄────────────────────►│ ComfyUI  │
│  (AI Agent)  │    command/prompt       │  (Python)    │   localhost:8188      │ (Server) │
└──────────────┘                         └──────────────┘                       └──────────┘

The MCP server translates natural language commands from the AI agent into ComfyUI API calls, then streams results back.

License

MIT — see LICENSE

Credits

Original project: ConstantineB6/comfy-pilot

Available Tools

14 tools
center_on_nodeA

Center the user's viewport on a specific node. Useful after creating nodes to show the user where they were placed.

Inputs: node_id (required).

download_modelA

Download a model to the ComfyUI models folder. Supports Hugging Face, CivitAI, and direct URLs. For gated HF models, will return instructions to provide a token.

Inputs: url (required), model_type (required), filename, hf_token, +1 more.

edit_graphA

Edit workflow graph with batched operations. Actions: create, delete, move, resize, set, connect, disconnect. Operations execute in order; 'create' returns node_id for chaining.

Inputs: operations (required).

get_node_infoB

Get detailed info about a specific node in the workflow: type, properties, inputs, outputs, widget values.

Inputs: node_id (required).

get_node_typesA

Search available node types. Returns minimal info by default. Use 'fields' for more details.

Inputs: fields, search, category.

get_statusA

Get ComfyUI status: queue, system stats, and/or history. Returns lightweight summaries by default (counts, IDs). Use detail='full' for more info. History is always paginated.

Inputs: detail, include, history_limit, history_offset.

get_workflowA

Get the current workflow from ComfyUI. Returns full node graph with all nodes, connections, and widget values. Use summarize_workflow for a lighter overview.

Inputs: no named parameters.

install_custom_nodeA

Install a custom node via ComfyUI Manager. Requires restart to complete.

Inputs: node_id (required).

runC

Run workflow or interrupt current generation.

Inputs: action, node_ids.

search_custom_nodesA

Search for custom nodes in the ComfyUI Manager registry. Returns name, author, description, install status, and star count.

Inputs: limit, query, status, category.

summarize_workflowA

Get a concise summary of the current workflow: node IDs, types, titles, positions, and connections. Lighter than get_workflow.

Inputs: no named parameters.

uninstall_custom_nodeA

Uninstall a custom node via ComfyUI Manager. Requires restart to complete.

Inputs: node_id (required).

update_custom_nodeB

Update a custom node to the latest version via ComfyUI Manager. Requires restart to complete.

Inputs: node_id (required).

view_imageA

View an image from a Preview Image or Save Image node. Returns the image as base64 so you can see it. Run the workflow first to generate images.

Inputs: node_id, image_index.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    A3.9/5.0

    Scored across 14 tools

    Disambiguation5/5

    Each tool serves a distinct purpose: workflow editing, node inspection, custom node management, model downloading, execution, and image viewing. There is no overlap or ambiguity.

    Naming Consistency5/5

    All tool names follow a consistent lowercase verb_noun pattern (e.g., download_model, get_workflow, search_custom_nodes), with clear verbs indicating the action.

    Tool Count5/5

    14 tools is well-scoped for a ComfyUI interface, covering essential operations like editing, custom node management, model downloading, and execution without being excessive.

    Completeness4/5

    The tool set covers core workflows (edit, run, view) and custom node lifecycle, but lacks explicit workflow save/load and queue management, which are minor gaps.

    Resources

    Related MCP Connectors

    Related MCP Servers

    • comfyui-mcp

      KerbalTheGathering
      F
      license
      Not graded
      quality
      D
      maintenance
      MCP server exposing a local (or LAN) ComfyUI instance's HTTP API as tools, so an LLM client can queue generations, inspect the queue/history, upload reference images, browse installed models/nodes, and pull back generated images.
      -

    • comfy-mcp

      cuzelac
      A
      license
      Not graded
      quality
      D
      maintenance
      MCP server that enables AI agents to control a local ComfyUI instance for image generation, allowing workflow understanding, parameter modification, execution, and model discovery.
      16 npm
      3
      Apache 2.0