MCP Integration

FFmpeg Micro MCP Server

Control video processing directly from Claude Code, Cursor, and any MCP-compatible AI tool. No scripts, no SDKs — just ask your AI assistant.

What is MCP?

The Model Context Protocol gives AI assistants secure, structured access to external tools.

MCP (Model Context Protocol) is an open standard that lets AI assistants call external APIs as “tools.” Instead of copying and pasting cURL commands or writing scripts, you describe what you want in plain language and your AI agent handles the rest.

FFmpeg Micro's MCP server exposes six tools that cover the full transcode lifecycle — from submitting jobs to downloading results. Your AI assistant authenticates via OAuth, so your API keys stay secure.

Quick Setup

Add the MCP server to your tool in under a minute.

Claude Code / Cursor
Add this to your project's .mcp.json
{
  "mcpServers": {
    "ffmpeg-micro": {
      "type": "http",
      "url": "https://mcp.ffmpeg-micro.com"
    }
  }
}
Authentication
OAuth 2.1 with PKCE — handled automatically

The first time your AI tool connects, it will open a browser window to sign in with your FFmpeg Micro account. After you approve, the token is cached — you won't be asked again.

No API keys to copy or environment variables to set. The MCP client and server negotiate OAuth automatically using dynamic client registration and PKCE.

Available Tools

Six tools covering the full transcode lifecycle.

transcode_video

Create a transcode job with full FFmpeg options, presets, or virtual effects. Returns immediately with a job ID.

transcode_and_wait

One-shot convenience tool: creates a job, polls until complete, and returns the result with a signed download URL.

get_transcode

Fetch the current state of a job by ID — status, output URL, processing time, and error details.

list_transcodes

List your transcode jobs with optional filters for status, time range, and pagination.

get_download_url

Generate a short-lived signed HTTPS URL for a completed job's output file.

cancel_transcode

Cancel a queued or in-progress transcode job.

Example Prompts

Just describe what you want — the AI agent picks the right tool.

Crop a video to square

Crop this video to a square: https://example.com/landscape.mp4

The AI agent calls transcode_and_wait with a crop filter and returns the download link.

Add a text overlay

Add the title "Episode 12" to the top of my video

Uses the @text-overlay virtual option to render styled text without hand-writing FFmpeg filters.

Batch convert to WebM

Convert all three of these MP4s to WebM at 720p

The agent creates multiple transcode jobs in parallel and monitors each to completion.

Compatible Tools
Any MCP-compatible AI client works.
  • Claude Code (CLI)
  • Claude Desktop
  • Cursor
  • Windsurf
  • VS Code (Copilot MCP)
API Documentation
Prefer REST? Use the API directly.

The MCP server calls the same REST API under the hood. Everything you can do via MCP is also available via direct HTTP requests.

View API docs →
Need Help?
• Browse support articles
• Check the dashboard
• Email: javid@ffmpeg-micro.com