DaVinci Resolve MCP Server 🎬🤖

A high-performance Model Context Protocol (MCP) server that connects DaVinci Resolve to your favorite local AI assistants—including Claude Desktop, Claude Code, Cursor, Windsurf, and Antigravity.

Control video editing, media pools, timelines, Fusion graphics, Fairlight audio, color grading, and render queues using natural language over standard I/O (stdio).

flowchart LR
    subgraph AI Client
        A["Claude Desktop / Cursor / Claude Code"]
    end

    subgraph MCP Server
        B["DaVinciResolveMCP (stdio)"]
        D["11 Embedded KB Resources"]
        E["49 Production Tools"]
    end

    subgraph Local Engine
        C["DaVinci Resolve (Studio/Native API)"]
    end

    A <-->|"JSON-RPC (stdio)"| B
    B --- D
    B --- E
    B <-->|"Python Scripting Bridge"| C

📑 Table of Contents

• 🌟 Features

• 📋 Prerequisites

• ⚙️ DaVinci Resolve Configuration

• 🚀 Quickstart Installation

• 🔌 Client Configuration

  • Claude Desktop

  • Claude Code CLI

  • Cursor / Windsurf / Antigravity

• 💬 Example Prompts & Use Cases

• 🧰 Tool Catalog (49 Granular Tools)

• 📚 Knowledge Base Resources (resolve-kb://)

• ⚡ Pre-Built Workflow Prompts

• 🧪 Testing Your Connection

• ❓ Troubleshooting & FAQ

• 🔒 Privacy & Security

• 📄 License

Related MCP server: DaVinci Resolve MCP Server

🌟 Features

• 🚀 100% Local & Fast (stdio): Communicates directly over standard I/O with zero open network ports, zero latency, and zero cloud lock-in.

• 🎬 49 Granular Tools: Deep coverage across all DaVinci Resolve pages:

  • Project & Database: Create, open, save, query settings, and inspect projects.

  • Media Pool: Ingest assets, organize folders/bins, and construct timelines.

  • Timeline & Inspector: Crop, pan, tilt, zoom, composite modes, track inspection, and colored markers.

  • Fusion & Titles: Text+, lower thirds, generators, and Fusion composition node inspection.

  • Fairlight Audio: AI Voice Isolation (0–100%), auto speech-to-text subtitles, and Fairlight track presets.

  • Color Grading: CDL balance (Slope/Offset/Power/Saturation), 3D LUTs (.cube), .drx grades, and color versions.

  • Deliver & Render: Export presets (YouTube, ProRes, H.264), queue management, render monitoring, and controls.

• 🧠 11 Embedded Knowledge Base Resources (resolve-kb://): Feeds the LLM comprehensive knowledge of Resolve API parameters, valid ranges, and safe workarounds.

• 🛡️ Defensive & Non-Destructive: Built-in timeline duplication, color version branching, and clear error diagnostics if Resolve is closed or misconfigured.

• 🌐 Cross-Platform: Works seamlessly on macOS, Windows, and Linux.

📋 Prerequisites

Before installing the server, ensure you have:

1. DaVinci Resolve:

   • DaVinci Resolve Studio is recommended for external scripting via the Python API. (Free version allows internal console scripts, but external automation requires Studio on most systems).

   • Compatible with DaVinci Resolve 18, 19, and 20+.

2. Python 3.10+:

   • Python 3.10, 3.11, 3.12, or 3.13 installed on your system.

3. An MCP-Compatible Client:

   • Claude Desktop, Claude Code, Cursor, Windsurf, Antigravity, or any standard MCP client.

⚙️ DaVinci Resolve Configuration

You must enable external scripting in DaVinci Resolve:

1. Launch DaVinci Resolve.

2. Open Preferences:

   • macOS: DaVinci Resolve > Preferences... (or Cmd + ,)

   • Windows / Linux: Edit > Preferences... (or Ctrl + ,)

3. Navigate to System > General.

4. Find the External scripting using setting and change it to Local (or Network).

5. Click Save.

6. Restart DaVinci Resolve if prompted.

🚀 Quickstart Installation

1. Clone the Repository

git clone https://github.com/au-yong/davinci-resolve-mcp.git
cd davinci-resolve-mcp

(Replace with your actual cloned path or current folder)

2. Set Up a Virtual Environment

# Create a virtual environment
python3 -m venv .venv

# Activate it:
# On macOS / Linux:
source .venv/bin/activate

# On Windows (PowerShell):
# .venv\Scripts\Activate.ps1

# On Windows (Command Prompt):
# .venv\Scripts\activate.bat

3. Install Dependencies

pip install -r requirements.txt

4. (Optional) Custom Install Paths via .env

If DaVinci Resolve is installed in default system locations, you don't need any extra configuration! The bridge automatically detects standard paths on macOS, Windows, and Linux.

If your installation is in a custom folder, copy .env.example to .env and set your paths:

cp .env.example .env

Edit .env according to your platform:

# Example for custom installation paths:
# RESOLVE_SCRIPT_API="/path/to/Developer/Scripting"
# RESOLVE_SCRIPT_LIB="/path/to/fusionscript.so"

🔌 Client Configuration

Connect the MCP server to your preferred client by adding the stdio command configuration.

1. Claude Desktop

Edit your Claude Desktop configuration file:

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

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

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

Add the davinci-resolve entry under mcpServers:

macOS / Linux Configuration:

{
  "mcpServers": {
    "davinci-resolve": {
      "command": "/ABSOLUTE/PATH/TO/DavinciResolveMCP/.venv/bin/python",
      "args": [
        "/ABSOLUTE/PATH/TO/DavinciResolveMCP/server.py"
      ],
      "env": {
        "PYTHONUNBUFFERED": "1"
      }
    }
  }
}

Windows Configuration:

{
  "mcpServers": {
    "davinci-resolve": {
      "command": "C:\\ABSOLUTE\\PATH\\TO\\DavinciResolveMCP\\.venv\\Scripts\\python.exe",
      "args": [
        "C:\\ABSOLUTE\\PATH\\TO\\DavinciResolveMCP\\server.py"
      ],
      "env": {
        "PYTHONUNBUFFERED": "1"
      }
    }
  }
}

Restart Claude Desktop. You will see the hammer 🔨 icon with 49 available DaVinci Resolve tools!

2. Claude Code CLI

You can register the server with Claude Code using the claude mcp add command:

# macOS / Linux
claude mcp add davinci-resolve /ABSOLUTE/PATH/TO/DavinciResolveMCP/.venv/bin/python /ABSOLUTE/PATH/TO/DavinciResolveMCP/server.py

# Windows
claude mcp add davinci-resolve C:\ABSOLUTE\PATH\TO\DavinciResolveMCP\.venv\Scripts\python.exe C:\ABSOLUTE\PATH\TO\DavinciResolveMCP\server.py

Or add it directly to your .claude.json / workspace configuration.

3. Cursor / Windsurf / Antigravity IDE

In your IDE's MCP Settings (e.g., Cursor Settings > Features > MCP, or Windsurf Settings > MCP), click Add New MCP Server:

• Name: davinci-resolve

• Type: stdio

• Command: /ABSOLUTE/PATH/TO/DavinciResolveMCP/.venv/bin/python

• Args: /ABSOLUTE/PATH/TO/DavinciResolveMCP/server.py

💬 Example Prompts & Use Cases

Once connected, you can talk to your AI assistant like an assistant video editor. Here are real examples of what you can ask:

🎬 Project Setup & Ingestion

"Check if DaVinci Resolve is connected, create a 4K 24fps project called 'Summer Travel Film', create a bin named 'B-Roll', and import all MP4 files from /Users/alex/Footage/Day1."

✂️ Rough Cut & Assembly

"Create a new timeline named 'Rough_Cut_v1', append the clips from the 'A-Roll' bin, and add a green marker at the start of each clip."

📱 9:16 Vertical Reel Repurposing

"Duplicate my active timeline to 'TikTok_9x16', change the resolution to 1080x1920, and adjust the scale and framing on all video track 1 clips so the speaker stays centered."

🎙️ Audio Cleaning & Auto-Subtitles

"Turn on AI Voice Isolation at 70% on all dialogue clips on Audio Track 1, then generate auto-subtitles from the dialogue audio."

🎨 Color Look Application

"Create a new color grade version called 'Cinematic Warm' on the first clip, apply the LUT from /Users/alex/LUTs/TealOrange.cube, and boost the saturation in the CDL."

🚀 Batch Export

"Load the YouTube 1080p render preset, set the output folder to /Users/alex/Exports, queue the render job, and start rendering."

🧰 Tool Catalog (49 Granular Tools)

1. Project Management (8 tools)

2. Media Pool & Bins (7 tools)

3. Timeline & Clip Inspector (10 tools)

4. Fusion, Titles & Graphics (5 tools)

5. Fairlight Audio & AI (5 tools)

6. Color Grading (5 tools)

7. Render & Export / Deliver Page (9 tools)

📚 Knowledge Base Resources (resolve-kb://)

The server exposes 11 embedded knowledge base resources. When your AI assistant needs to verify exact parameter limits, supported codecs, or script workarounds, it can automatically consult these resources:

• resolve-kb://capabilities: Capability tiers (Direct API vs Workarounds vs UI-only)

• resolve-kb://effects: Resolve FX / OpenFX parameters and recommended baseline values

• resolve-kb://transitions: Supported video and audio transition identifiers

• resolve-kb://graphics: Title insertion templates, generators, and Text+ recipes

• resolve-kb://fusion: Fusion node graphs, tool IDs, and keyframe scripting

• resolve-kb://audio: Fairlight EQ, dynamics, LUFS standards, and Voice Isolation

• resolve-kb://color: Node trees, Rec.709 balancing, CDL math, LUTs, and DRX grades

• resolve-kb://inspector: Complete SetProperty keys, value types, and valid ranges

• resolve-kb://recipes: Step-by-step production recipes (YouTube, social clips, podcast)

• resolve-kb://readme: Knowledge base golden rules and API boundaries

• resolve-kb://api-reference: Raw DaVinci Resolve Scripting API reference extract

⚡ Pre-Built Workflow Prompts

The server includes pre-configured workflow prompts that can be triggered directly in your client:

• youtube_talking_head_workflow: Orchestrates project creation, 1080p setup, A-Roll bin ingestion, timeline creation, Voice Isolation at 65%, speech-to-text subtitles, and a YouTube export queue job.

• vertical_social_reel_workflow: Automates duplicating a landscape timeline, converting resolution to 9:16 (1080x1920), adjusting Pan/Tilt/Zoom framing, and generating open captions.

🧪 Testing Your Connection

You can verify that everything is working properly from your terminal:

# Activate your virtual environment first
source .venv/bin/activate  # Or Windows equivalent

# Run diagnostic check
python -c "
import server
print('Connection Status:', server.get_resolve_status())
"

If successful, you will see output similar to:

Connection Status: {
    'connected': True,
    'product_name': 'DaVinci Resolve Studio',
    'version': '19.1.0.0020',
    'current_project': 'My Project',
    'current_timeline': 'Timeline 1',
    'timeline_count': 1
}

❓ Troubleshooting & FAQ

1. "Could not connect to DaVinci Resolve"

• Is DaVinci Resolve open? Resolve must be running with a project loaded before running tools.

• Is External Scripting enabled? Go to Preferences > System > General and make sure External scripting using is set to Local or Network.

• Studio vs Free: External Python scripting is primarily supported in DaVinci Resolve Studio. The free version restricts external scripting processes on some operating systems.

2. "ModuleNotFoundError: No module named 'DaVinciResolveScript'"

The bridge automatically searches default operating system directories:

• macOS: /Library/Application Support/Blackmagic Design/DaVinci Resolve/Developer/Scripting

• Windows: C:\ProgramData\Blackmagic Design\DaVinci Resolve\Support\Developer\Scripting

• Linux: /opt/resolve/Developer/Scripting

If your Resolve installation is non-standard, define RESOLVE_SCRIPT_API and RESOLVE_SCRIPT_LIB in your .env file (see .env.example).

3. Claude Desktop doesn't show the tools

• Check that the paths in claude_desktop_config.json are absolute paths (no ~ or relative paths).

• Ensure you used the path to the python executable inside your virtual environment (.venv/bin/python or .venv\Scripts\python.exe).

• Check Claude Desktop logs at:

  • macOS: ~/Library/Logs/Claude/mcp*.log

  • Windows: %APPDATA%\Claude\logs

🔒 Privacy & Security

• 100% Local Execution: The server communicates exclusively via standard input/output (stdio).

• No External Ports: No HTTP or WebSocket ports are opened on your machine.

• No Telemetry: No logs, media files, project data, or personal information are collected or transmitted anywhere.

📄 License

This project is licensed under the MIT License.

Happy Editing with AI! 🎬✨ If you find this project helpful, please give it a star on GitHub!