resolve-mcp

The most comprehensive MCP server for DaVinci Resolve. 295+ tools covering the complete DaVinci Resolve scripting API (v20.3) — control every aspect of Resolve from Claude, ChatGPT, or any MCP-compatible AI assistant.

What is resolve-mcp?

resolve-mcp is a Model Context Protocol (MCP) server that exposes 295+ tools for controlling DaVinci Resolve through AI assistants like Claude, Claude Code, Cursor, Windsurf, ChatGPT, and any other MCP client.

It turns natural language into Resolve scripting API calls — letting you manage projects, edit timelines, grade color, render deliverables, manage media pools, control Fusion compositions, and more, all through conversation with an AI.

Why resolve-mcp?

• Complete API coverage — 295+ tools spanning every Resolve scripting API surface: projects, media pool, timelines, editing, markers, rendering, color grading, Fusion, Fairlight, galleries, node graphs, Dolby Vision, and stereo 3D

• Zero code required — just talk to your AI assistant: "Create a new timeline called 'Final Cut' at 24fps" or "Add a dissolve transition to every cut point"

• One-line install — uvx resolve-mcp or pip install resolve-mcp

• Works with any MCP client — Claude Desktop, Claude Code, Cursor, Windsurf, VS Code + Continue, or your own MCP client

• Optional AI editing tools — 3 bonus Gemini-powered tools for AI timeline analysis, marker-driven editing, and automated marker placement

• Cross-platform — macOS, Windows, Linux (wherever Resolve runs)

Related MCP server: DaVinci MCP Professional

Quick Start

Prerequisites

1. DaVinci Resolve (Free or Studio) must be running

2. Enable scripting: Preferences → System → General → External scripting using = Network

3. Python 3.11+

4. ffmpeg (for AI tools only): brew install ffmpeg (macOS) or apt install ffmpeg (Linux)

Install

# Using uvx (recommended — runs in isolated environment)
uvx resolve-mcp

# Using pip
pip install resolve-mcp

# Using pipx
pipx install resolve-mcp

# Install with the AI editing assistant (resolve-assistant) included
pip install resolve-mcp[assistant]

Configure Claude Desktop

Add to your Claude Desktop MCP config:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

• Windows: %APPDATA%\Claude\claude_desktop_config.json

• Linux: ~/.config/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "resolve-mcp": {
      "command": "uvx",
      "args": ["resolve-mcp"]
    }
  }
}

Configure Claude Code

Add to your project's .mcp.json:

{
  "mcpServers": {
    "resolve-mcp": {
      "command": "uvx",
      "args": ["resolve-mcp"]
    }
  }
}

Configure Cursor / Windsurf / VS Code

Follow your editor's MCP server configuration docs, pointing to uvx resolve-mcp as the command.

Enable AI Bridge Tools (Optional)

To use the 3 Gemini-powered AI tools (resolve_analyze_timeline, resolve_add_markers, resolve_build_from_markers), set your Gemini API key:

{
  "mcpServers": {
    "resolve-mcp": {
      "command": "uvx",
      "args": ["resolve-mcp"],
      "env": {
        "GEMINI_API_KEY": "your-gemini-api-key-here"
      }
    }
  }
}

Windows note: If uvx is not on your system PATH, use the full path to uvx.exe (e.g., C:\\Users\\YOU\\.local\\bin\\uvx.exe) as the "command" value.

Get a free Gemini API key at aistudio.google.com.

Using Both Servers Together — resolve-mcp + resolve-assistant

resolve-mcp gives you 295+ tools for direct DaVinci Resolve control — projects, timelines, editing, color grading, rendering, Fusion, Fairlight, and more.

resolve-assistant is a separate MCP server that adds AI-powered automatic editing — point it at a folder of footage and give it an editing instruction, and it uses Google Gemini to watch every frame, plan the edit, and build the timeline in Resolve automatically.

They are independent MCP servers. You can use either one alone, or both together for the complete AI video editing toolkit. When used together, you get full manual control over Resolve (resolve-mcp) plus AI-driven footage analysis and automatic timeline building (resolve-assistant).

Install Both With One Command

# Install resolve-mcp with the AI editing assistant included
pip install resolve-mcp[assistant]

# Or install them separately
pip install resolve-mcp
pip install resolve-assistant

Configure Both Servers

Because they are two separate MCP servers, each needs its own entry in your Claude Desktop or Claude Code config. Here is the complete configuration with both servers:

Claude Desktop (~/Library/Application Support/Claude/claude_desktop_config.json on macOS, %APPDATA%\Claude\claude_desktop_config.json on Windows, ~/.config/Claude/claude_desktop_config.json on Linux):

{
  "mcpServers": {
    "resolve-mcp": {
      "command": "uvx",
      "args": ["resolve-mcp"],
      "env": {
        "GEMINI_API_KEY": "your-gemini-api-key-here"
      }
    },
    "resolve-assistant": {
      "command": "uvx",
      "args": ["resolve-assistant"],
      "env": {
        "GEMINI_API_KEY": "your-gemini-api-key-here"
      }
    }
  }
}

Claude Code (.mcp.json in your project root):

{
  "mcpServers": {
    "resolve-mcp": {
      "command": "uvx",
      "args": ["resolve-mcp"],
      "env": {
        "GEMINI_API_KEY": "your-gemini-api-key-here"
      }
    },
    "resolve-assistant": {
      "command": "uvx",
      "args": ["resolve-assistant"],
      "env": {
        "GEMINI_API_KEY": "your-gemini-api-key-here"
      }
    }
  }
}

Note: GEMINI_API_KEY is optional for resolve-mcp (only the 3 AI bridge tools need it) but required for resolve-assistant. Both servers use the same Gemini API key.

Which Server Do I Need?

For more about the AI editing assistant, see the resolve-assistant documentation.

Tool Categories

Project Management (10 tools)

Create, open, close, save, and configure DaVinci Resolve projects. Switch databases, manage project settings, and query project metadata.

Media Pool (21 tools)

Import, organize, search, and manage clips in the media pool. Create bins, move clips, set metadata, relink media, and more.

Timeline Management (15 tools)

Create, duplicate, delete, import, and export timelines. Manage tracks, set timeline properties, and navigate between timelines.

Editing (12 tools)

Manipulate timeline items — set properties, transforms, composite modes, speed changes, and clip-level operations.

Markers & Playhead (9 tools)

Add, edit, delete, and query markers on timelines. Control playhead position.

Rendering (14 tools)

Configure render settings, manage the render queue, and control rendering jobs.

Color Grading (12 tools)

Control color grading — apply LUTs, set CDL values, manage nodes, copy grades, grab stills, and export frames.

Fusion (8 tools)

Manage Fusion compositions on timeline clips.

Fairlight Audio (4 tools)

Control Fairlight audio features — voice isolation, audio insertion, and track info.

Additional Categories

• Clip Metadata (18 tools) — markers, flags, proxy management, transcription

• Timeline Items (11 tools) — item-level markers and flags

• Clip Versions (11 tools) — version management, color group assignment

• Gallery & Stills (7 tools) — still albums, still export/import

• Node Graph (5 tools) — node labels, enable/disable, tool inspection

• Layout Presets (22 tools) — UI layouts, render presets, burn-in presets

• Dolby Vision & Stereo 3D (4 tools) — Dolby Vision analysis, stereo conversion

• Media Storage (3 tools) — browse volumes and storage locations

• Project Manager (10 tools) — archive, delete, import, restore projects

• Bin/Folder Operations (4 tools) — transcription, metadata export, folder IDs

• Generators & Titles (13 tools) — insert generators, titles, Fusion compositions, scene cut detection

AI Bridge Tools (3 tools, requires GEMINI_API_KEY)

MCP Resources (Read-Only)

Usage Examples

Once configured, just talk to your AI assistant naturally:

Project Management

"List all my Resolve projects" "Create a new project called 'Summer Campaign' at 4K 24fps" "What are the current project settings?"

Media Pool

"Import all files from /Volumes/Media/Footage into the 'Raw' bin" "How many clips are in the 'B-Roll' bin?" "Set the clip color of all clips in 'Interviews' to Blue"

Timeline Editing

"Create a new timeline called 'Final Cut v3' with 2 video and 4 audio tracks" "List all clips on video track 1" "Add a Blue marker at frame 1200 with the note 'Great reaction shot'"

Rendering

"Set up a render job for H.265 MP4 at 4K" "Start rendering and tell me when it's done" "What render presets are available?"

Color Grading

"Apply the 'Kodak 2383' LUT to the current clip" "Set the CDL lift to add warmth to the shadows" "Grab a still from the current frame"

AI-Powered Analysis (requires Gemini)

"Analyze the current timeline and give me editorial feedback" "Build a timeline from markers using the footage in /Volumes/Media/Footage"

Claude Code Plugin — Skills, Agents & Hooks

This repo is also a Claude Code plugin that installs 13 editor-focused skills, 3 specialized agents, and safety hooks directly into Claude Code.

Install Everything With One Command

claude plugin add github.com/jenkinsm13/resolve-mcp

This installs the MCP server (295+ tools) AND all skills, agents, and hooks in one step. No separate MCP configuration needed — the plugin bundles it all.

Note: Set your GEMINI_API_KEY in the MCP server config after install if you want the AI bridge tools. The 215+ Resolve tools work without it.

Skills (invoke with /skill-name)

Agents

Hooks

• .env protection — blocks editing of .env files containing API keys

• Destructive operation warning — warns before delete operations in Resolve (clips, timelines, grades, markers, bins, projects)

Note: The plugin bundles both the MCP server and all skills/agents/hooks. For standalone MCP server install (without skills), use uvx resolve-mcp or pip install resolve-mcp.

Architecture

resolve-mcp
├── resolve_mcp/               # Python package (installed via pip/uvx)
│   ├── __init__.py            # Package init, registers all tool modules
│   ├── __main__.py            # python -m resolve_mcp entry point
│   ├── config.py              # FastMCP server + optional Gemini client
│   ├── resolve.py             # DaVinci Resolve scripting API connection
│   ├── errors.py              # Error handling + @safe_resolve_call decorator
│   ├── resources.py           # MCP resources (resolve://project, etc.)
│   ├── project_tools.py       # Project management (10 tools)
│   ├── media_pool_tools.py    # Media pool operations (11 tools)
│   ├── timeline_mgmt_tools.py # Timeline management (15 tools)
│   ├── edit_tools.py          # Timeline item editing (12 tools)
│   ├── marker_tools.py        # Markers & playhead (9 tools)
│   ├── render_tools.py        # Render pipeline (14 tools)
│   ├── color_tools.py         # Color grading (12 tools)
│   ├── fusion_tools.py        # Fusion comp management (8 tools)
│   ├── fairlight_tools.py     # Fairlight audio (4 tools)
│   ├── ...                    # 10+ more tool modules
│   └── resolve_ai_tools.py    # AI bridge tools (3 tools)
├── skills/                    # Claude Code plugin skills
│   ├── color-assist/SKILL.md
│   ├── deliver/SKILL.md
│   ├── preflight/SKILL.md
│   └── ...                    # 13 skills total
├── agents/                    # Claude Code plugin agents
│   ├── color-analyst.md
│   ├── timeline-auditor.md
│   └── color-consistency-checker.md
├── hooks/hooks.json           # Claude Code plugin hooks
├── .claude-plugin/plugin.json # Plugin manifest
├── pyproject.toml
├── .env.example
└── README.md

How It Works

1. resolve-mcp starts a FastMCP server that registers 295+ tools

2. Your AI client (Claude, Cursor, etc.) connects via the MCP protocol

3. When you make a request, the AI selects the appropriate tool(s)

4. Each tool calls the DaVinci Resolve scripting API via the Python SDK

5. Results are returned as structured text to the AI for interpretation

The server connects to Resolve over the network scripting interface, which must be enabled in Resolve's preferences. No plugins or extensions needed — just enable the scripting API and run the server.

Troubleshooting

"DaVinci Resolve is not running"

• Make sure Resolve is open

• Enable scripting: Preferences → System → General → External scripting using = Network

• Restart Resolve after enabling scripting

"No project open in Resolve"

• Open or create a project in Resolve before using project-dependent tools

Tools not appearing in Claude Desktop

• Restart Claude Desktop after modifying claude_desktop_config.json

• Check the MCP server logs in Claude Desktop's developer console

• Verify uvx resolve-mcp runs successfully from the terminal

AI bridge tools return "GEMINI_API_KEY not set"

• Add GEMINI_API_KEY to the env section of your MCP config

• The 215+ Resolve tools work without Gemini — only the 3 AI bridge tools require it

DaVinci Resolve API Compatibility

• Tested with: DaVinci Resolve 20.3 (Free and Studio)

• Platform support: macOS (Apple Silicon + Intel), Windows, Linux

• Studio-only features are clearly marked — the server gracefully handles Free edition limitations

Related Projects

• resolve-assistant — AI-powered video editing assistant that uses Google Gemini to analyze footage, plan professional edits, and build timelines in DaVinci Resolve automatically. Install alongside resolve-mcp with pip install resolve-mcp[assistant] for the complete AI video editing toolkit.

• FastMCP — The MCP framework powering this server

• Model Context Protocol — The open protocol for AI tool use

License

MIT License. See LICENSE for details.

Contributing

Contributions welcome! Please open an issue or PR on GitHub.

Built for video editors, colorists, and post-production professionals who want to control DaVinci Resolve with AI. Works with Claude Desktop, Claude Code, Cursor, Windsurf, ChatGPT, and any MCP-compatible client. Pair with resolve-assistant for AI-powered automatic editing.