get_video_contentLABSCO SUMMARY
One call from a video link to text, with four speech vendors behind the same interface.
The provider indirection is the useful part: you configure whichever keys you already have and the server picks in a fixed order — Deepgram first, then Gladia, Speechmatics, AssemblyAI — so switching vendors is a key change, not a code change. Transcription is not free either way, so the cost of a long video lands on your account with that provider, not here.
What it is
A Python server that takes a video link, extracts the audio, and returns the transcribed text through get_video_content.
What you get
- Audio pulled from YouTube, Bilibili, TikTok, Twitter and the other sites yt-dlp covers — over a thousand of them
- Transcription through Deepgram, Gladia, Speechmatics or AssemblyAI
- Provider chosen automatically from whichever API keys you configured, in that order
- Speaker separation, so a two-person interview does not come back as one block
- Async handling, with temporary files cleaned up after the job
Requirements
At least one transcription key in .env — DEEPGRAM_API_KEY, GLADIA_API_KEY, SPEECHMATICS_API_KEY or ASSEMBLYAI_API_KEY. Python 3.11 or higher is recommended, installed with uv. Served over SSE on port 8000 by default. YouTube may ask for authentication; the fix is a cookies.txt in the project root.
Setup effort
Build from source — git clone https://github.com/ruezo/mcp-video-digest.git
USE CASES
01
Getting the substance of an hour-long talk without watching it
Hand over the URL and read the transcript, then ask follow-up questions against it in the same conversation.
02
Pulling quotes out of an interview
Speaker separation keeps the turns apart, so you can tell who said the line you want to quote.
03
Working with Chinese-language video
The repository points at Deepgram for Chinese content, and the provider order puts it first when the key is present.
04
Transcribing something behind a YouTube sign-in wall
Export your browser cookies to cookies.txt in the project root and the download step stops failing.
1 tool
Listed in project docs. Your client may show a different list.
BEFORE YOU INSTALL
What to know first
Tools you get
1 tool named in the project's documentation.
Requires
- uv (or Python)
- An API key for at least one transcription service: DEEPGRAM_API_KEY, GLADIA_API_KEY, SPEECHMATICS_API_KEY or ASSEMBLYAI_API_KEY, in the project's .env file
INSTALLATION
Connect your client
This one is built from source rather than installed with a single command.
Server setup
TERMINAL
git clone https://github.com/ruezo/mcp-video-digest.git
cd mcp-video-digest
uv venv
uv pip install -e .Rename .env.example to .env and add the API key for at least one transcription service — DEEPGRAM_API_KEY, GLADIA_API_KEY, SPEECHMATICS_API_KEY or ASSEMBLYAI_API_KEY
Start the server and leave it running (it serves SSE on port 8000, the address in the client block):
TERMINAL
uv run src/main.pyMCP Clients
Any MCP client — SSE
ANY MCP CLIENT — SSE · JSON
{
"mcpServers": {
"video_digest": {
"url": "http://<YOUR_IP>:8000/sse"
}
}
}<YOUR_IP> is the address of the machine running uv run src/main.py — localhost if it is your own.
