DaVinci Resolve Local Bridge MCP

What DaVinci Resolve Local Bridge MCP does

DaVinci Resolve Local Bridge MCP MCP server connects an MCP-compatible AI assistant to a running DaVinci Resolve installation. It exposes Resolve operations as MCP tools, covering application state, project management, project settings, timelines, media, Fusion, color grading, rendering, and supported optional extensions. The README describes 218 available actions, each surfaced with a resolve_ prefix after discovery.

The server supports both DaVinci Resolve Free and Studio. Its in-process bridge approach is particularly relevant to the Free edition, where external scripting is restricted. The bridge runs within Resolve and uses the live Resolve scripting object, allowing read and write access through the available API actions.

How it works

DaVinci Resolve Local Bridge MCP MCP server uses two local Python processes. The MCP server communicates with an AI client over standard input and output, then sends requests over HTTP to 127.0.0.1:8787. A bridge script running inside Resolve receives those requests and executes named actions.

The bridge provides health, action-listing, and action-execution endpoints. At startup, the MCP server queries the action list and creates corresponding MCP tools automatically, so tool definitions do not need to be maintained separately. Calls are sent to the bridge with a parameter object. The bridge remains active through a blocking HTTP server until stopped or replaced.

The normal launcher is a Workflow Integration inside Resolve. The bridge can also be started through the Resolve Scripts area as a fallback. A previous process using the configured port is terminated when a new bridge starts.

Setup and configuration

Requirements are DaVinci Resolve 18.5 or newer, Python 3.10 or newer, and an MCP-compatible client. Run python install.py to install the Workflow Integration and bridge files, then restart Resolve so the integration is detected. The dependency installation command is pip install -r requirements.txt.

Add mcp_server/server.py to the client configuration using a standard stdio server entry. The README gives configurations for Claude Desktop, Cursor, Devin, and Claude Code, and the server can also be launched directly with python mcp_server/server.py.

After opening Resolve, choose Workspace → Workflow Integrations → DaVinci Resolve Local Bridge MCP Server Connection and select Start Bridge. The panel should report the local HTTP address and the number of available actions. The Resolve Workspace Console contains connection and startup diagnostics.

Manual installation requires copying the workflow integration files and bridge script to Resolve's platform-specific script locations. The README lists paths for Windows, macOS, and Linux.

Tools and capabilities

The exposed actions include tools for:

• Checking bridge connectivity, Resolve version, current page, project, and timeline state.
• Creating, loading, saving, closing, deleting, importing, exporting, and archiving projects.
• Managing project folders, databases, project names, and project settings.
• Switching between Resolve pages such as Edit, Cut, Fusion, Color, Fairlight, Deliver, and Media.
• Working with timelines, media, Fusion, color grading, rendering, and supported optional extensions.

The documented status entry point is resolve_status, which reports bridge connectivity and current project or timeline information. Other actions follow the same resolve_ naming pattern and accept the parameters defined by their discovered action.

Limitations and notes

The bridge must be running inside an open Resolve session before the MCP tools can perform Resolve operations. Communication is local and uses port 8787 on the loopback address; the provided material does not describe remote access or authentication.

The available tool set is discovered from the bridge rather than hard-coded in the MCP server. Consequently, the actions reported by the running bridge are the source of truth for the current session. Resolve 18.5 or newer and Python 3.10 or newer are required according to the setup instructions.

The README identifies the bridge as having full read and write access to the Resolve API. AI-generated changes can therefore alter projects and media-related state, so clients should be configured and prompted with appropriate care before executing mutating actions.