// Transcription MCP Server

Give your AI agent a media processing tool.

Install a transcription MCP starter so agents can clean, compress, transcribe, subtitle, summarize, and download agent-ready transcript JSON from audio or video.

MCPnpxJSON

MCP starter ↓CLI starter ↓Skill starter ↓

See pricing Full product

vocce · transcribe● live

// runs in your agent, ships to your stack Claude Code Cursor Gemini MCP CLI REST API n8n Zapier Make GitHub Actions Notion HubSpot

// What you get

One upload. Every file the next step needs.

The same reliable Vocce pipeline, focused on this job. Start free, then plans from $9.90/mo when you need more.

create_transcription_job

get_transcription_job

list_transcription_artifacts

agent.v1 schema

Webhook delivery

// How it works

How to give your AI agent transcription

Upload or paste a URL Any format, any length — Vocce normalizes it.

One reliable call Clean, compress, transcribe, diarize, summarize.

Export the pack Transcript, subtitles, summary, and agent JSON.

// who uses transcription mcp server

Built for real workflows.

Claude & MCP agents

One `claude mcp add` and your agent can clean, transcribe, subtitle, and summarize any media a user hands it.

Agent products

Ship voice-note and meeting features without building a media pipeline — your agent calls one stable tool.

Automation builders

Call it from an n8n node or a GitHub Action today, with Zapier and Make coming. Same tool names everywhere, so flows stay portable.

// faq

Transcription MCP Server, answered.

What is a transcription MCP server? +

An MCP (Model Context Protocol) server exposes transcription as tools an AI agent can call directly — create_transcription_job, get_transcription_job — with a stable schema the agent can rely on.

How do I install it? +

One command: claude mcp add vocce -- npx -y @vocce/transcribe-mcp. Set your API key in the environment and the tools appear in your agent.

What does the agent get back? +

agent.v1 JSON: speaker turns, timestamps, chapters, summaries, quality warnings, and artifact links — structured data an agent can reason over, not a wall of text.

Do failed jobs burn credits? +

No — jobs are idempotent with automatic retries, and only successful completions are billed.

// related tools

More from the same engine.

Automate transcription from the terminal. Give agents transcripts they can actually reason over. Convert audio to text, then export the whole pack. Turn interviews into clean, quotable transcripts. Whisper-quality transcription without the setup. Turn video into transcripts, captions, and reusable notes.