Skip to content
MCP server

EarScribe MCP

Free browser-based audio-to-text transcription with timestamps and subtitle export

Pricing
Free
Platforms
Web, API, self-hosted
Version
0.1.0
Listing updated
today

Similar MCP servers

About EarScribe MCP

EarScribe MCP is a free MCP server for design and media. It is available on Web, API and self-hosted.

The README describes EarScribe MCP as a read-only Model Context Protocol server that exposes EarScribe's local-first guidance, FAQ and official links to AI clients such as Claude Desktop, Cursor, Windsurf and Continue. It does not itself transcribe audio; the transcription tool is the EarScribe web app, which the server documents.

The underlying app runs Whisper models in the browser through WebGPU or WebAssembly, so audio stays on the user's device. The site lists MP3, WAV, M4A, OGG, FLAC and WebM input, video containers for audio extraction, and output as TXT, SRT, VTT or JSON.

The server's two tools return setup guidance and official links as markdown. It is aimed at people who want their MCP client to answer questions about EarScribe, or who want a quick reference for running the transcription workflow locally.

Tools this server provides (2)

Taken from the project documentation; the running server may differ.

get_local_setup
Return canonical local-setup guidance for running the AI workflow on-device.
get_official_links
Return the canonical list of official links for EarScribe.

What it does

On-device transcription
Audio is processed in the browser with Whisper, so recordings are not uploaded to external servers, according to the site.
Timestamped transcript editing
Transcript lines link to positions in the recording and can be corrected before export.
Subtitle export formats
Output can be downloaded as SRT, VTT, TXT or JSON.
Automatic language detection
Whisper detects the spoken language across the 99 supported languages without manual selection.
Selectable model sizes
Users can pick Whisper models from Tiny to Turbo to trade download size against accuracy.
Read-only MCP resources
The server exposes setup notes, an FAQ and official links to MCP clients without requiring API keys.

Good for

  • Generate subtitle files from a recorded video or podcast episode
  • Transcribe interviews that should stay on a local machine
  • Turn lecture recordings into searchable notes
  • Ask an MCP-compatible AI client for EarScribe setup steps

Pricing

Free

Web app free with no account; MCP server is MIT licensed

Prices change. Check earscribe.app before you buy.

EarScribe MCP FAQ

Does the MCP server transcribe audio?

The README describes the server as exposing EarScribe knowledge, FAQ and official links. Transcription happens in the EarScribe web app.

What tools does the MCP server provide?

It provides get_local_setup, which returns local-setup guidance, and get_official_links, which returns official links. Both take no parameters.

Is there a cost?

The website states that EarScribe is free with no account, and the README lists the MIT license for the server repository.

Which export formats are available?

The site and README list TXT, SRT, VTT and JSON exports.

Is EarScribe MCP free?

Yes. EarScribe MCP is free to use.

What are the best alternatives to EarScribe MCP?

The closest alternatives to EarScribe MCP are WordCast MCP, Speechmatics MCP by usefulapi, Audio Transcription MCP, Audio & Video Transcriber (Whisper) and Whipscribe MCP.

Where this comes from

  • Description: written from the project's README, the project's website and its MCP Registry entry.

How AIPoweredBot checks servers