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., "@Reel transcribe ~/Movies/team-standup.mp4 with hints: Reel, utt"
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.
An MCP server that turns a video into text, on this Mac.
Point it at any video ffmpeg can read. It extracts the sound, splits it on silence when the video is long, has utt transcribe each piece on-device, and hands the stitched transcript back to whichever assistant asked. Reel also appears as a settings page inside utt, listing the videos it has transcribed.
ffmpeg and ffprobe (brew install ffmpeg)
utt, running
Related MCP server: video-to-llm
bun install
just install-page # draw the settings page inside utt
just register # make the tools available to every Claude Code sessionTool | What it does |
|
|
|
|
hints are words a recogniser tends to get wrong — names, products, repos.
Near misses are corrected to them.
Let assistants turn video into text — the switch that turns every request away.
Keep a list of the videos — what the page shows.
Save the words beside the video — leaves a .txt next to each file.
Command | Description |
| Format, lint, typecheck, tests |
| Write the plugin manifest into utt |
| Add the MCP server to every Claude Code session |
This server cannot be deployed
Transform video, audio and images, and generate media from prompts. FFmpeg, captions, models.
FFmpeg as a service for AI agents: typed video editing tools, async jobs, downloadable outputs.
Timestamped transcripts, chapters and clip suggestions from YouTube, Twitch, Kick or TikTok links.
YouTube transcripts, search, channels, playlists and bulk transcript jobs for AI agents. 14 tools.
