Kdenlive MCP
Provides a headless video editor that uses Kdenlive for review, with a portable JSON project as source of truth and compilable MLT XML.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Kdenlive MCPrender my video project to mp4"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Kdenlive MCP
Local-first video editing through the Model Context Protocol. Kdenlive MCP keeps a versioned JSON project as the source of truth, applies revision-checked timeline edits, compiles deterministic MLT, queues renders, generates previews, and verifies outputs using the Kdenlive/MLT and FFmpeg installation already on your computer.
Public preview: Milestones 0-4, 6, and 7 are implemented. Local AI analysis (Milestone 5) remains planned and unavailable.
What works
Canonical projects with atomic revisions, checkpoints, undo, and redo.
Media ingest with allowed-root containment and managed/external modes.
Clip placement, linked A/V, ripple edits, properties, keyframes, effects, transitions, text, captions, and markers.
Durable preview/export jobs, progress, cancellation, restart recovery, and bounded artifacts.
Output duration, stream, A/V sync, loudness, clipping, black, and freeze checks.
MCP over stdio or authenticated loopback Streamable HTTP.
Automatic Kdenlive/MLT/FFmpeg discovery and an actionable doctor command.
Editable Kdenlive 26.04.x and OTIO 0.18.1 export, plus reviewed import plans.
Versioned editing recipes and advisory, durable timeline diagnostics.
Related MCP server: Cutible MCP Server
Quick start
Requirements: Node.js 22+, pnpm 9, and Kdenlive or separate MLT/FFmpeg tools.
Powershell:
git clone https://github.com/AMMIROSOH/Kdenlive-mcp.git
cd Kdenlive-mcp
Set-ExecutionPolicy -Scope Process Bypass
.\scripts\install.ps1 -AddToPath
kdenlive-mcp --doctorLinux:
git clone https://github.com/AMMIROSOH/Kdenlive-mcp.git
cd Kdenlive-mcp
./scripts/install.sh
~/.local/share/kdenlive-mcp/kdenlive-mcp-*/kdenlive-mcp --doctorThen add the stdio command to your MCP client with an absolute allowed root:
kdenlive-mcp --root /path/to/video-workspace --client-id my-mcp-clientSee the complete installation and client configuration guide.
Safety model
Files are limited to configured roots; symlink/traversal escapes are rejected.
Mutations require the exact current revision and remain undoable per client.
Render processes use argument arrays, no shell, bounded output/time, and a restricted environment.
HTTP binds only to loopback and requires a generated bearer token.
Kdenlive, MLT, and FFmpeg binaries are not redistributed by this repository.
Read the threat model and security policy before exposing the server to additional software.
Documentation
Development
pnpm install --frozen-lockfile
pnpm check
pnpm doctor
pnpm render:acceptance
pnpm mcp:acceptance
pnpm docs:check
pnpm sbom
pnpm release:acceptance -- --platform windowsLicense
Apache-2.0 for this repository. External Kdenlive/MLT/FFmpeg installations retain their own licenses; see third-party notices.
This server cannot be deployed
Maintenance
Related MCP Connectors
A real timeline video editor for AI agents: journaled edits, FFmpeg/MLT rendering, exports
Render video and run AI media tasks from a single declarative JSON request.
- VidmoatOAuthcom.vidmoat
AI video editor: create projects, edit timelines, add captions and effects, and render videos.
Generate, edit, and render videos via the Shotstack video editing API.
Related MCP Servers
- FlicenseAqualityBmaintenanceEnables AI-powered video editing in IDEs through a pipeline-driven workflow, including proxy ingestion, motion graphics generation, and high-fidelity rendering.4-
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to perform headless video editing through 35 tools for project creation, clip manipulation, rendering, quality control, and semantic search, all via JSON-RPC 2.0 over stdio.1MIT
- FlicenseNot gradedqualityBmaintenanceLocal-first AI video editing engine controlled by MCP-compatible AI hosts, enabling AI to inspect footage, propose edits, modify multitrack projects, and render locally while keeping user control.-
- AlicenseNot gradedqualityBmaintenanceA real video editor for AI agents, served over MCP, enabling journaled timeline editing, rendering via FFmpeg/MLT, and deterministic CLI operation.1MIT