Figma Context MCP
Give your coding agent access to your Figma data. Implement designs in any framework in one-shot.
Descripción del creador (original en inglés)
Professional MCP server for video editing automation with OpenCut.
Estos tres nunca se combinan en una sola puntuación: las opiniones, las mediciones y la popularidad son cosas distintas.
Todavía no escribimos un resumen sencillo para este. Abajo están los datos que leímos del registro.
Elige la app de IA que usas y copia la configuración.
Necesitas
¿Qué app usas?
Elige tu app arriba. La recordamos, así que la próxima vez el botón dirá “Instalar en …”.
Abre Claude Desktop → Settings → Developer → Edit Config y pega esto en claude_desktop_config.json
{
"mcpServers": {
"opencut-controller": {
"command": "npx",
"args": [
"-y",
"opencut-controller"
]
}
}
}Ejecuta esto en tu terminal
claude mcp add opencut-controller -- npx -y opencut-controllerPégalo en .cursor/mcp.json de tu proyecto, o en ~/.cursor/mcp.json
{
"mcpServers": {
"opencut-controller": {
"command": "npx",
"args": [
"-y",
"opencut-controller"
]
}
}
}Tu navegador te pide permiso para abrir la app; luego la app muestra la configuración y te pide confirmar. Los valores de las claves quedan vacíos para que los completes tú.
Pégalo en .vscode/mcp.json de tu proyecto
{
"servers": {
"opencut-controller": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"opencut-controller"
]
}
}
}Tu navegador te pide permiso para abrir la app; luego la app muestra la configuración y te pide confirmar. Los valores de las claves quedan vacíos para que los completes tú.
Este se ejecuta en tu computadora, y ChatGPT solo puede conectarse a servidores en internet. Elige otra app.
Tu elección se guarda en una cookie en este dispositivo; no necesitas cuenta.
Esta configuración se genera a partir de la ficha del creador en el registro. Si el README dice otra cosa, sigue el README.
Se ejecuta en tu computadora; la lista de herramientas aparece en tu app de IA después de instalarlo.
Revisamos información que cualquiera puede ver; no hemos leído el código. Las revisiones que no aplican no se cuentan. · Revisado el 6 nov 2026
El nombre en el registro se verificó con la cuenta de GitHub JXUE0, dueña del repo
Todavía no hemos leído este repo en GitHub
Todavía no hemos leído este repo en GitHub
Este se ejecuta en tu computadora; no hay endpoint en internet
Se ejecuta en tu computadora y usa ajustes locales en su lugar
Se ejecuta en tu computadora; solo podemos leer listas de herramientas de endpoints en internet
Comparado con todos los servidores activos de esta categoría en el registro. Las estrellas solo cuentan cuando quien lo publica es dueño del repo.
Superó 1 de 1 revisiones básicas aplicables. No hemos leído el código. Antes de agregar una clave o un token, revisa qué acceso pide el servidor y usa una clave que tenga solo los permisos que necesitas.
El registro no indica ninguna clave obligatoria, aunque algunos servidores te piden iniciar sesión la primera vez que los usas.
Se ejecuta en tu computadora. Tu app de IA lo inicia cuando hace falta, y los datos no pasan por los servidores del creador (a menos que el propio servidor llame a un servicio en línea).
El repo de GitHub cambió por última vez –, y la última versión (1.30.0) llegó al registro hace 6 meses.
Las apps compatibles con MCP, como Claude Desktop, Claude Code, Cursor y VS Code. También necesitas Node.js en tu computadora.
En la misma categoría: Figma Context MCP, Claude Real Video, Figma MCP Server
Servidores de la misma categoría hechos por otros creadores, ordenados por estrellas.
Cómo leer una tarjeta:★ Estrellas de GitHubiTres pilares: reseñas · revisiones · usoiActualizado = fecha en que se publicó la última versión en el registroi
Give your coding agent access to your Figma data. Implement designs in any framework in one-shot.
Let any LLM watch a video locally — and search everything it has ever watched.
The Figma MCP server brings Figma design context directly into your AI workflow.
Copiado tal cual del repo del creador. PopNx MCP no lo ha revisado ni lo respalda. Se quitaron HTML, imágenes y scripts. Los enlaces llevan fuera de este sitio.
🌐 Leer en Español: [README.es.md](./README.es.md)
[!NOTE]
OpenCut Controller is a robust Model Context Protocol (MCP) server designed to fully automate and control the OpenCut Video Editor directly from AI models (like Claude or agents inside n8n).
By leveraging Playwright under the hood, this server seamlessly injects into the OpenCut browser environment, granting LLMs programmatic access to manipulate the timeline, scenes, media assets, rendering engine, and more—effectively turning OpenCut into an AI-driven, headless video editing powerhouse.
stdio (default) or integrate remotely via modern HTTP Streamable transport.Ensure you have pnpm (or Bun) installed on your system.
[!IMPORTANT]
OpenCut Controller requires the Legacy OpenCut (Next.js App Router) — the version with a complete editor, timeline, and stores.
The current public repo (
OpenCut-app/OpenCut) is an early rewrite (Rust core + GPUI desktop app + Vite web shell) whose editor is not implemented yet, so there is nothing to control there.
Use the archived legacy repo:
OpenCut-app/opencut-classic— upstream itself recommends it for now.
# 1. Clone the legacy OpenCut editor (required for MCP to work)
git clone https://github.com/OpenCut-app/opencut-classic.git cd opencut-classic
cd .. git clone https://github.com/JXUE0/opencut-controller.git cd opencut-controller
pnpm install
pnpm run patch-editor ../opencut-classic
cd ../opencut-classic bun install cd apps/web && bunx next dev --turbopack -p 3001
[!TIP]
The
pnpm installcommand will automatically trigger apostinstallscript to download the required Playwright Chromium binaries. No manual browser setup is needed!
[!NOTE]
The patcher is idempotent: running it again on an already patched editor is a no-op. If you patched with a version older than 1.32.0, restore
apps/web/src/components/providers/editor-provider.tsxfrom git before re-running it — the old patch exposed store state instead of the stores.
| Variable | Default | Description |
|---|---|---|
OPENCUT_URL | http://localhost:3001 | URL of the patched local OpenCut editor. |
MCP_HEADLESS | true | Set to false to watch the browser while the MCP works. |
CHROMIUM_PATH | — | Use a specific Chromium/Chrome binary instead of Playwright's. |
TRANSPORT_TYPE | stdio | stdio or http. |
PORT | 3002 | HTTP transport port. |
MCP_AUTH_TOKEN | — | Bearer token for the HTTP transport (strongly recommended). |
Conventions: every time value (startTime, duration, time, trimStart, …) is in seconds; the controller converts to OpenCut's internal ticks. FPS accepts plain numbers (30, 29.97). Element volume is in dB (-60 to 20, 0 = original).
The server can be launched using two different transport protocols depending on your client environment.
stdio Transport (Default)Ideal for standard local MCP clients like Claude Desktop.
npx -y opencut-controller
Latest version straight from GitHub (built on install): npx -y --allow-git=all github:JXUE0/opencut-controller (--allow-git is required since npm 12). From a clone of this repo: pnpm build && node dist/index.js.
HTTP Streamable TransportIdeal for integrating with external workflow tools like n8n or remote services. The server listens at http://localhost:3002/mcp.
[!WARNING]
Environment variable syntax varies depending on your operating system. Make sure to use the correct command below to prevent errors.
Linux / macOS (Bash):
TRANSPORT_TYPE=http PORT=3002 npx -y opencut-controller
Windows (PowerShell):
$env:TRANSPORT_TYPE="http"; $env:PORT="3002"; npx -y opencut-controller
For secure remote access, set a bearer token:
MCP_AUTH_TOKEN=your-secure-token TRANSPORT_TYPE=http PORT=3002 npx -y opencut-controller
Then include Authorization: Bearer your-secure-token in requests.
The controller exposes 161 highly granular tools to the LLM, categorized as follows:
| Category | Tools | Description |
|---|---|---|
| 📁 Project | 16 | Create, load, save, duplicate projects and change their settings. |
| 🎬 Scenes | 6 | Manage scene lifecycle, renaming, and active state. |
| ▶️ Playback | 10 | Play, pause, stop, seek, volume, and mute. |
| ⏱️ Timeline Tracks & Settings | 10 | List, inspect, mute, hide tracks; timeline settings. |
| 🧩 Timeline Elements | 15 | Insert, trim, split, move, duplicate, and update clips. |
| ✨ Timeline Effects | 5 | Apply and tweak clip effects. |
| 📌 Keyframes | 6 | Add, retime, and remove animation keyframes. |
| 🎯 Selection/Clipboard | 6 | Select elements, copy and paste. |
| 🕰️ History | 4 | Undo and redo. |
| 🎞️ Media | 7 | Import media files and place them on the timeline. |
| 📝 Text | 9 | Create and style text overlays. |
| 🎵 Audio | 8 | Search music/SFX and set element volume. |
| 🎨 Stickers/Canvas | 13 | Add stickers; canvas size, FPS, and background. |
| ⚙️ Transcribe/Export | 11 | Captions from transcripts; export to MP4/WebM files. |
| 🔖 Bookmarks/Panels/Keybindings | 23 | Timeline bookmarks, UI panels, and shortcuts. |
| 🔑 Auth & Storage | 9 | User sessions and storage. |
| 🌐 API | 3 | Health check and sound search via OpenCut's backend. |
[!NOTE]
OpenCut's main track is magnetic: its first clip always starts at
0. Use an overlay element (text, sticker, a second clip) for free positioning. Tracks are created automatically when an element is inserted withouttrackId; OpenCut deletes empty tracks, sotimeline_track_addreturns an explanatory error.
export_project accepts outputPath to write the rendered .mp4/.webm file to disk.
Resources provide context directly to the LLM about the current state of OpenCut:
opencut://projects - JSON array of all available OpenCut projects.opencut://editor/state - Live tracking of the active project, scene, playhead time, and playback status.opencut://timeline/tracks - Detailed breakdown of tracks in the active scene.Quick-start templates for complex actions:
create_intro_video - Automates creating a 10-second intro with text overlays.add_background_music - Searches the OpenCut sound library and trims music to fit.apply_transition - Seamlessly blends two clips with a specified effect.To connect OpenCut Controller to your local Claude Desktop app, add the following to your claude_desktop_config.json:
{
"mcpServers": { "opencut-controller": { "command": "npx", "args": ["-y", "opencut-controller"], "env": { "OPENCUT_URL": "http://localhost:3001" } } } }
The browser is closed automatically when the client disconnects.
postinstall ran successfully. You can trigger it manually with npx playwright install chromium. npm 12+ blocks install scripts by default, so with npx run npx playwright install chromium once before first use.$env:TRANSPORT_TYPE="http"; npx -y opencut-controller instead.OPENCUT_URL (default port 3001, not Next.js' default 3000).[!TIP]
To check types during development run
pnpm run typecheck;pnpm run buildbundlesdist/.
window.__opencut is undefined / Editor not ready after 30000msCause: You are trying to use the MCP server with the production web app (opencut.app). Fix: You must run the patched OpenCut editor locally (pnpm run patch-editor ). The production site doesn't expose the internal hooks required for MCP control. If the editor fails to compile, check its terminal for module-not-found errors.
MP4 export forces software H.264 encoding (--disable-accelerated-video-encode) because the Windows hardware encoder stalls OpenCut's exporter, and runs in Chromium's new headless mode because the legacy headless shell crashes while rendering.
Incompatible React versionsCause: react and react-dom versions mismatch in the monorepo. Fix: Ensure both are set to the exact same version (e.g., 19.2.0) in the root package.json and run pnpm install.
Invalid input: expected string, received undefined (Zod Error)Cause: Missing environment variables in apps/web/.env. Fix: Ensure you have a .env file in apps/web/. We have updated the code to make most variables optional, but BETTER_AUTH_SECRET may still be needed for some features.
Cause: The local OpenCut server (Next.js) is still compiling. Fix: Wait until you see ✓ Ready in the editor terminal before starting the MCP server.
This project's original codebase and concept were inspired by and built upon the foundational work done in RavenMeld/OpenCut-MCP. We extend our sincere gratitude to RavenMeld for their pioneering work on connecting OpenCut via the Model Context Protocol.
This project is licensed under the MIT License.