DaVinci Resolve MCP Server

DaVinci Resolve MCP Server: A Thorough Guide to the Model Context Protocol for AI-enabled Resolve Workflows

[Version badge] [npm badge] [API Coverage badge] [Tools badge] [Advanced badge] [Tested badge] [DaVinci Resolve badge] [Python badge] [License badge]

A Model Context Protocol (MCP) server that lets AI assistants control DaVinci Resolve Studio through the official Scripting API. It provides full API coverage plus guarded workflow helpers for editing, media pool organization, render setup, review markers, grading, Fusion, Fairlight, project lifecycle tasks, extension authoring, and source-safe media analysis. A local browser control panel ships with the server for inspecting Resolve state, running source-safe analysis, drilling into analyzed clips and shots, and editing analysis output inline. See the Control Panel Guide for the full tour.

Introduction and visual context

• This MCP server is designed to bridge AI agents with DaVinci Resolve Studio, unlocking scripted control over the Resolve environment while maintaining safeguards and auditability. It offers extensive API coverage and task-specific tools to accelerate post-production workflows.
• A local control panel is included to visualize Resolve state, inspect analyzed media, and adjust analysis output live. The panel provides a browsable, in-browser view of clips, shots, and analysis results, enabling hands-on validation alongside automated workflows.
• The project emphasizes source-safe media analysis and a robust set of tooling for editing, media organization, grading, and delivery planning, all while keeping source media immutable unless the user explicitly requests changes.

Quick Start: how to begin

• Quick setup command
• npx davinci-resolve-mcp setup
• Before connecting to Resolve, configure external scripting in Resolve:
• Preferences > General > External scripting using → Local
• What happens during installation
• The npm launcher installs a managed copy under your user application-data directory and runs a universal Python installer.
• The installer creates a Python virtual environment, detects Resolve paths, and can configure third-party tooling (e.g., Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, Zed, Continue, Cline, Roo Code, OpenCode, and JetBrains IDEs).
• For source installs
• git clone https://github.com/samuelgursky/davinci-resolve-mcp.git
• cd davinci-resolve-mcp
• python install.py
• Additional setup notes
• For device-specific paths and manual configuration, consult Installation and Configuration.
• The installer checks for the latest MCP release and can prompt, snooze, or opt into auto-updates. Checks are best-effort and non-blocking to startup.
• Documentation and deeper setup
• The Installation and Configuration guide includes platform paths, client config, and manual setup steps.
• The installer can manage multiple MCP entries (compound and advanced) in a single install.

Launching the Local Control Panel: a hands-on look

• Start the control panel with a single-user approach
• venv/bin/python -m src.control_panel
• What you’ll see
• A local web interface that lets you inspect Resolve state, drill into analyzed media, and adjust outputs inline.
• The panel is a practical companion for AI agents, letting you verify decisions or tweak analysis results side-by-side with Resolve’s UI.
• When to use this panel
• Use it to validate analysis outputs, confirm markers and metadata, and cross-check how AI-driven decisions map to actual Resolve timeline content.

Server Modes: choosing how you want the MCP server to operate

• Compound mode (default)
• Entry point: src/server.py
• Tools: 34 tools, forming a comprehensive surface that groups related Resolve operations behind action parameters to minimize context switching.
• Best for most AI assistants, offering a balance of capability and workflow cohesion.
• Full / granular mode
• Entry point: src/server.py --full or src/resolvemcpserver.py
• Tools: 341 tools
• Best for power users who want one MCP tool per Resolve API method, enabling ultra-fine-grained control.
• Advanced server (Node-based, optional)
• Package: davinci-resolve-advanced-mcp
• Binary: bin/davinci-resolve-advanced-mcp.mjs
• What it does
  • Reads and edits Resolve files (.drp, .drt, .drx) and applies DB/XML-level changes without running Resolve, enabling cloud or local operation without Resolve GUI.
  • Includes 18 tools spanning post-production domains: drp, drt, drx per-clip grading, offline QC catalogs, color science capabilities, fusion, audio, Fairlight, media, editorial, and provenance tooling.
  • Live-calibrated DRX grade writes against Resolve Studio units; programmatic cleanup capabilities for node graphs and project databases.
• The advanced server can also be used as a library (importable engine API), not just as a running server.
• A notable feature: the DB-as-truth pipeline that compiles YAML project specs into a canonical SQLite DB and runs stages with governance gates, provenance, and drift detection.
• The core is pure JavaScript with MIT licensing; some features require external tools (e.g., ffmpeg for audio, sharp or better-sqlite3 for certain paths).
• Both the standard and advanced servers are installable together within a single npm install, with a shared configuration that enables both to run concurrently.
• Bradford Post Assistant — an optional, closed-beta companion
• Desktop application on top of the MCP foundation
• A copilot that watches Resolve sessions (timeline, grades, frames) with embedded AI agent runtime and local media analysis
• Persistent on-device memory, cross-episode learning, and a self-contained design that handles updates automatically
• Focused on professional post workflows: grading, conform QC, delivery validation, and Fusion composition authoring
• Access is invitation-based via Bradford Operations; more information at Bradford Operations’ site
• The MCP ecosystem is designed so you can run the live Resolve control while optionally using offline or advanced tooling for deeper workflow automation or offline grade authoring.

What you can do with MCP: practical examples

• Example prompts and actions you can issue to the MCP
• List all projects and open the one called “My Film”
• Create a timeline called “Assembly Cut” from all clips in the current bin
• Build a multicam prep timeline from selected camera angles and preserve source media
• Detect 2-pops or slate claps and suggest record offsets for sync prep
• Publish analysis summaries, keywords, people, and slate hints into Resolve clip metadata
• Probe this timeline for gaps, overlaps, missing media, and source frame ranges
• Safely import an image sequence, organize it into bins, and normalize clip metadata
• Build a ProRes 422 HQ render plan, validate the settings, and queue the job
• Copy review markers from the timeline to the selected clip and export a review report
• Snapshot a clip’s grade, validate a CDL update, and export a temp LUT
• Create a Fusion TextPlus overlay on the selected clip and verify graph connections
• Report audio channel mappings, voice isolation availability, and subtitle support
• Install MCP-marked DCTL or script, classify refresh/restart needs, then remove it
• Core capabilities: what the compound server covers
• App and project control: Launch/reconnect, project CRUD, folders, databases, cloud wrappers, settings, presets, archives
• Media pool and ingest: Safe import, image sequences, multicam prep, bin organization, metadata normalization and inventory, marks and annotations, relinking/proxy handling
• Media analysis: Source-safe file/clip/bin/project analysis, 2-pop/slate-clap detection, metadata writes, persisted analysis artifacts, host-chat visual analysis, transcription
• Timeline editing and conform: Probing, text edits, copy/move/duplicate, range ops, gaps/overlaps, source ranges
• Review annotations: Markers, data, flags, color, cleanup, review reports
• Color and grading: Node graph probing, CDL validation, DRX/LUT helpers, versions, gallery stills
• Fusion: Timeline item comps, safe tool creation, port inspection, bulk writes
• Audio and Fairlight: Probing, mapping, guarded writes, transcription, voice isolation
• Render and deliver: Format/codec support, render settings validation, queue lifecycle checks
• Extension authoring: MCP-bound helper workflows for Fuse, DCTL, ACES DCTL, and Resolve-page scripting
• Source media safety and security
• Media safety: Treats originals as immutable; analysis writes to sidecars or project analyses only; metadata publishing goes to Resolve’s project database
• Security posture: Local stdio process, no network listener, no multi-user auth surface by default
• The security policy document details operational boundaries, confirmation guidance, and vulnerability reporting

Key stats and documentation: what to expect from the project

• Key metrics
• MCP Tools: 34 compound tools, 341 granular tools
• Advanced tools: 18 offline tools (drp/drt/drx, DB authoring, and more)
• Kernel actions: 136 guarded workflow actions across nine compound tools
• API methods covered: 336 of 336 (100%)
• Live-tested methods: 331 of 336 (98.5%)
• Live test pass rate: 331/331 (100%)
• Tested against: DaVinci Resolve 19.1.3 Studio + Resolve 20.3.2 Studio
• Documentation and guides
• Installation and Configuration: System requirements and setup options
• API Coverage and Test Results: Stats, coverage, and test status
• Kernel Action Coverage: The current guarded workflow map
• AI Skill Reference: Guidance for AI assistants using the compound server
• Control Panel Guide: Local browser panel tour (Overview, Analyze, Setup, Preferences)
• Media Analysis Guide: Source-safe FFprobe/FFmpeg/Whisper workflows and analysis roots
• Multicam Setup Helper Guide: Timeline prep and UI conversion steps
• Editorial Decision Guide: Editorial craft guidance for analysis and timelines
• Color Decision Guide: Project-owned color guidance and Resolve color API boundaries
• Contributing and Project Layout: Contribution workflow and repository structure
• Security Policy: Local stdio trust boundaries
• Release Process: Maintainer checklist and release notes
• Changelog: Historical notes
• Extensions and notes
• Extension authoring references live in docs/authoring
• Developer-package notes live in docs/notes and docs/integrations
• Prompt recipes live in examples

Requirements and development: what you need to run MCP

• DaVinci Resolve Studio 18.5+ (macOS, Windows, or Linux)
• Note: The free edition does not support external scripting
• Python 3.10+ (3.10–3.12 is the lowest-risk range; 3.13/3.14 work on newer Resolve builds)
• Resolve 19.1.3 remains the compatibility baseline; Resolve 20.x adds capabilities
• Resolve 21+ features are behind runtime capability detection and activate automatically on newer builds
• Resolve external scripting set to Local
• Development commands
• Compound server
  • python src/server.py
• Granular server
  • python src/server.py --full
• Quick local tests
  • venv/bin/python tests/test_import.py
  • venv/bin/python scripts/auditapiparity.py
• Release and validation
• Documentation at docs/process/release-process.md
• AI agents guidance in AGENTS.md
• Claude Code users can refer to CLAUDE.md

License, authors, and acknowledgments

• License: MIT
• Author: Samuel Gursky ([email protected])
• GitHub: github.com/samuelgursky
• Acknowledgments
• Blackmagic Design for DaVinci Resolve and its scripting API
• The Model Context Protocol team for enabling AI assistant integration

Images and visuals in the MCP post

• Control Panel image
• A local browser control panel ships with the server for inspecting Resolve state, analyzing data, and editing analysis output inline. The image illustrates the control panel’s overview and layout (docs/images/control-panel/01-overview.png).
• Badge strip and platform visuals
• Version badge: version-2.60.0
• npm badge: npm version
• API Coverage badge: 100% coverage
• Tools badge: 34 compound / 341 granular tools
• Advanced badge: 18 offline tools
• Live tested badge: 98.5% live-tested
• DaVinci Resolve badge: Resolve 18.5+ support
• Python badge: Python 3.10+
• License badge: MIT
• These badges visually summarize the MCP’s scope and status and appear at the top of the post.

A final note on usage and best practices

• Use the compound server for general-purpose AI-assisted workflows, where a broad capability surface is beneficial and context is managed automatically.
• Switch to the granular or advanced server when you need precise, method-level control, or when offline-grade control and DB-level edits are required without Resolve running.
• Leverage the Bradford Post Assistant if you want an on-device AI copilot for post workflows, with memory, learning, and a closed-beta integration path.
• Always follow the Security Policy: keep the server as a local, private tool unless you explicitly authorize networked or multi-user configurations.
• Consult the documentation sections for installation, API coverage, kernel actions, and the control panel tour to maximize the MCP’s potential.

This MCP server suite represents a mature, feature-rich approach to AI-assisted DaVinci Resolve workflows, offering a careful balance of safety, automation, and powerful tooling. Whether you’re building AI agents that can operate Resolve through its scripting API, or you’re exploring deeper, offline-grade manipulation with the advanced MCP, the platform is designed to scale with your post-production needs while keeping media safety and project integrity at the forefront.