gradio-transcript-mcp: A Gradio MCP Server for Audio/Video Trans…

@bsmnyk

Repository

Basic information

About gradio-transcript-mcp: A Gradio MCP Server for Audio/Video Transcription from URLs

Gradio demo cum MCP server to generate transcripts from Audio/Video

#gradio #transcription#audio-video

gradio-transcript-mcp: A Gradio MCP Server for Audio/Video Transcription from URLs config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "gradio-transcript-mcp": {
      "command": "python",
      "args": [
        "app.py"
      ]
    }
  }
}

gradio-transcript-mcp: A Gradio MCP Server for Audio/Video Transcription from URLs tools

Tool list not available yet

This listing does not currently have a tool list to display. This does not mean the server has no tools; check its documentation for supported capabilities.

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is gradio-transcript-mcp?

gradio-transcript-mcp is a Gradio application configured as an MCP (Model Control Protocol) server that transcribes audio and video from URLs into text. It uses OpenAI's Whisper and ffmpeg (via yt-dlp) to download, convert, and transcribe multimedia content, enabling MCP clients like Cline to process media inputs efficiently.

How to use gradio-transcript-mcp?

Clone the repository, install dependencies with pip install -r requirements.txt, and run python app.py to start the local Gradio web interface and MCP server. The server exposes a transcribe_url tool, accessible at http://127.0.0.1:7860/gradio_api/mcp/sse. For clients that don't support SSE directly, use mcp-remote as an intermediary. A hosted version is also available at https://bismay-gradio-transcript-mcp.hf.space/gradio_api/mcp/sse.

Key features of gradio-transcript-mcp

Transcribes audio/video from any URL into English text
Uses OpenAI's Whisper model via the transformers library
Converts media to WAV format before transcription
Dynamic device selection (CPU or GPU)
Supports local and hosted (Hugging Face Spaces) deployment
Exposes a single MCP tool: transcribe_url

Use cases of gradio-transcript-mcp

Transcribe a podcast, lecture, or meeting from a direct URL
Integrate multimedia transcription into MCP‑powered workflows (e.g., Cline)
Convert video content to text for indexing or analysis
Automate transcription of user‑submitted media links

Frequently asked questions

What does gradio-transcript-mcp transcribe?

It transcribes audio and video from a provided URL into English text. It downloads the media, converts it to WAV, and runs Whisper transcription.

What are the runtime dependencies?

Python, `gradio[mcp]`, `yt-dlp`, `transformers`, `torch`, and ffmpeg. The `ffmpeg_setup.py` script ensures ffmpeg is available.

Where does the transcribed data live?

Transcription happens on the server (local or Hugging Face Spaces). No data is stored permanently; the tool returns the transcription as a string.

How do I connect an MCP client?

Use the SSE URL printed when running the app (e.g., `http://127.0.0.1:7860/gradio_api/mcp/sse`). Configure your client with that URL. For clients not supporting SSE directly, use `npx mcp-remote` as an intermediary.

Does it require authentication?

No authentication is mentioned. The server runs locally or on a public Hugging Face Space without built‑in auth.

Comments

More Media & Design MCP servers

Generate Greetings

Create greetings and personalized PNG cards for 30 occasions in 15 languages, including Hindi, Greek and Hebrew. Free remote MCP with curated phrases, occasion-matched backgrounds and four styles, including motivational.

ProfessionalHeadshot.io

Official LinkedIn, Slack and GitHub profile photo specs with sources, headshot style guides by use and industry, and a free 0-100 headshot audit. No API key needed.

EmbedSocial

Connect Claude, ChatGPT, and other AI assistants to your social media content, reviews, and UGC aggregated via EmbedSocial.

diggercamp

Tommaso Savorana (ti es)

diggercamp finds independent music by how it sounds. Paste any Bandcamp, YouTube or SoundCloud link and get the closest sonic matches from 4M+ independent tracks — matched by rhythm, timbre, key and energy, not genre tag

Kinetune

Turn your songs into lyric videos (9:16, 16:9, 1:1) and Spotify Canvas loops. Find a song, browse Looks, get an exact credit quote, and render only after you say yes.