@audiolabtools/mcp-server

MCP (Model Context Protocol) server that gives any MCP-capable AI — Claude Desktop, Claude Code, Cursor, and others — nine audio-analysis tools, backed by the hosted AudioLab API. It is a thin HTTP client: no local audio engine, no ffmpeg, nothing to compile. It can analyse a public URL or a local file on your machine.

Install

Point your MCP client at the package via npx (nothing to install globally):

{
  "mcpServers": {
    "audiolab": {
      "command": "npx",
      "args": ["-y", "@audiolabtools/mcp-server"],
      "env": { "AUDIOLAB_API_KEY": "al_live_yourkey" }
    }
  }
}

Get a key: sign in at https://audiolab.tools/account and generate one (free tier available).

Related MCP server: Audio Analysis MCP Server

Requirements

• Node ≥ 18 — uses the built-in global fetch + AbortSignal.timeout.

• An AUDIOLAB_API_KEY. No ffmpeg, no native dependencies.

Tools

Every tool takes one audio source — a public url or a local path:

• { url: "https://…" } — a public https URL the API fetches server-side.

• { path: "./mix.wav" } — a file on the machine running this server. Files up to 4 MB are sent inline; larger files (up to 50 MB) upload over a one-shot signed URL, are analysed, and are then deleted. (Local path works only in this stdio server, not the remote /mcp endpoint.)

Example asks to your AI:

• “Analyze the loudness of https://example.com/track.wav” → analyze_loudness with url

• “Run loudness on ./master.wav” → analyze_loudness with path

• “Does ./mix.mp3 pass Spotify?” → check_target with path + target:"spotify"

Configuration (env)

Privacy

Analysis happens on the AudioLab API, so the audio does reach audiolab.tools — a url is fetched server-side, and a local path is sent to the API (small files inline; larger files via a private one-shot signed upload that is deleted right after analysis). The API returns numbers only and does not retain your audio (see https://audiolab.tools/privacy). This package has no telemetry and writes nothing to disk. If audio must never leave the machine, don't use a hosted analyser.

Limits

• Local files: up to 50 MB (host bigger ones at a public URL).

• Duration: loudness routes (analyze_loudness, check_target, analyze_timeseries, get_spectrum) handle long files (podcast episodes, full sets — up to ~3 h) via server-side streaming; voice/signal routes are limited to ~7 minutes.

• One file per call (agents loop for many); one-shot (no streaming/realtime).

• Rate and monthly limits are enforced by the API, per key.

Smoke test

node hosted-server.mjs --selftest   # verifies the 9 tools + guards; no network

License

MIT © Nathan Renting