agentleFS
umakhanproumakhanpro/vocalingo-mcp/skills/transcribe-audio/SKILL.md

transcribe-audio

Skill0 starsChanged 16 days agoMITView on GitHub

Turn a voice message, audio recording, call, interview or video soundtrack into text with the VocaLingo MCP server. Detects the language. Use when the user shares a voice note (Telegram, WhatsApp), an mp3/m4a/ogg/wav file or a recording and asks what was said, wants it written down or needs a transcript.

7 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Call speech_to_text { source }.

    line 13
  2. Very specific:

    Every error has a code and a nextStep: follow nextStep.

    line 35
  3. Very specific:

    If no VocaLingo tools are available, the server is not connected yet.

    line 39

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Transcribe a voice message or recording

Works through the VocaLingo MCP server (https://nodejs.vocalingo.app/mcp). Tools: speech_to_text, summarize_media, get_upload_url, check_job.

Steps

  1. Get the source: a direct file link (a Telegram Bot API file link works) or a fileId after upload (see "Files" below).
  2. Call speech_to_text { source }. It returns text, language and durationSeconds.
  3. Give the user the text. If they asked for the meaning in another language, translate the text yourself: no extra VocaLingo call is needed.

Pick summarize_media instead when the recording is long (a meeting, lecture or podcast), bigger than 50 MB, comes as a YouTube/TikTok/Instagram link, or the user wants a summary or timecodes. Its transcript is available through check_job { jobId, include: "transcript" }.

Cost

About 1 VocaLingo token per 2 minutes of audio. Files up to 50 MB.

Files

  • Public link: pass it as source.
  • Agent with a shell (Claude Code, Cursor, Codex, Gemini CLI): call get_upload_url { filename }, run the returned curlCommand with the local path, then pass the fileId as source. A fileId stays valid for 48 hours, so reuse it instead of uploading again.
  • Chat app (claude.ai, Claude apps): call get_upload_url and the user gets an upload button in the chat (if there is no button, give them browserUploadUrl). Wait for their message with the fileId. Do not upload from a code sandbox: its network blocks the upload domain.
  • ChatGPT: a file attached to the message reaches the tool through its file parameter, no upload step needed.

Tokens and confirmation

Tools spend VocaLingo tokens, the currency of the user's VocaLingo account (unrelated to LLM tokens). A call that would cost 50 tokens or more does not start: it returns status: "confirmation_required" with the price and a confirmToken. Show the price and ask. Only after the user agrees, repeat the same call with confirmToken added. Never confirm on the user's behalf. dryRun: true returns the estimate without starting anything. On insufficient_balance, tell the user to top up in the VocaLingo app and stop.

Errors

Every error has a code and a nextStep: follow nextStep. If the same call fails twice, show the error to the user and stop.

Not connected

If no VocaLingo tools are available, the server is not connected yet. Send the user to https://github.com/umakhanpro/vocalingo-mcp#connect. They need a VocaLingo account (https://vocalingo.app) and sign in through the browser.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.