Magic Master MCP — audio mastering for AI agents

MCP serverMedia

Your AI can master audio tracks to specific LUFS and True Peak loudness targets and remove AI fingerprints from Suno or Udio audio. mastering adds these audio finishing tools to your AI, so it can prepare tracks the way you need them.

Available today. Use it from your connected AI after setup.

Add mastering, then tell your AI which audio to work on and the loudness target you want to hit, or ask it to remove Suno or Udio AI fingerprints.

Then ask your AI: use Magic Master MCP — audio mastering for AI agents to get service info

What your AI can do with it

  • Master audio tracks to a chosen LUFS loudness target
  • Hit True Peak level targets on mastered audio
  • Remove AI fingerprints from Suno-generated audio
  • Remove AI fingerprints from Udio-generated audio
  • Prepare finished audio that meets your loudness targets

From the project's README

As published by denisok-ai/magicmaster-mcp in README.md.

Remote MCP server that lets Claude, ChatGPT, Cursor, VS Code, Claude Code and any other MCP client master audio: hit a LUFS target with True Peak limiting, strip the Suno / Udio AI fingerprint, and get a mastering passport back in the chat.

Nothing to install: the server is hosted. This repository holds the connection recipes and a tiny stdio bridge for clients that cannot speak HTTP.

One-click install

Connect

Claude Code

claude mcp add --transport http magicmaster https://magicmaster.pro/mcp

claude.ai / Claude Desktop (custom connector, OAuth): Settings → Connectors → Add custom connector → URL https://magicmaster.pro/mcp/oauth → sign in on the consent page. Use https://magicmaster.pro/mcp if you only need the free tools.

Claude Desktop (claude_desktop_config.json)

{ "mcpServers": { "magicmaster": { "type": "http", "url": "https://magicmaster.pro/mcp" } } }

Cursor (.cursor/mcp.json)

{ "mcpServers": { "magicmaster": { "url": "https://magicmaster.pro/mcp" } } }

VS Code (.vscode/mcp.json)

{ "servers": { "magicmaster": { "type": "http", "url": "https://magicmaster.pro/mcp" } } }

ChatGPT (Developer mode → Connectors → Create): URL https://magicmaster.pro/mcp/oauth, authentication OAuth.

stdio-only clients — bridge through mcp-remote:

{ "mcpServers": { "magicmaster": { "command": "npx", "args": ["-y", "mcp-remote", "https://magicmaster.pro/mcp/oauth"] } } }

or the npm package — npx -y magicmaster-mcp (magicmaster-mcp on npm), a one-line wrapper around the same bridge:

{ "mcpServers": { "magicmaster": { "command": "npx", "args": ["-y", "magicmaster-mcp"] } } }

Docker

For clients without Node, and for the Glama build that runs security checks:

docker build -t magicmaster-mcp .
docker run -i --rm magicmaster-mcp

The image holds only the bridge — the server itself stays hosted at https://magicmaster.pro/mcp. It is built from this repository (npm ci), so mcp-remote is baked in and the container needs no registry at run time — only outbound HTTPS to magicmaster.pro. The image defaults to the anonymous endpoint (free tools, no sign-in); switch it to the OAuth endpoint, which also unlocks the paid tools, with -e MAGICMASTER_MCP_URL=https://magicmaster.pro/mcp/oauth.

Build spec for a Glama release (the platform configures the recipe on its side, it does not read this Dockerfile):

  • build steps: npm ci --omit=dev
  • start command: node bin/magicmaster-mcp.js
  • environment: MAGICMASTER_MCP_URL=https://magicmaster.pro/mcp

Tools

ToolWhat it doesCost
analyze_trackLUFS, true peak, duration, correlation, genre hintfree
clean_ai_traceremoves the Suno / Udio digital fingerprint, loudness untouchedfree
master_trackmastering: 24 presets, LUFS target, 6 export formats1 token
get_jobstatus + result metrics, with a mastering passport widget (MCP Apps)free
check_limitsbalance, remaining free quota, reset timefree
list_presets, get_pricing, get_service_infopresets with target loudness, live prices, service manifestfree
create_topup_linkprepares a checkout for the human — the agent cannot pay by itselffree

Typical flow: analyze_track → (clean_ai_trace for AI-generated tracks) → master_track → poll get_job. A master takes 20–60 s.

Access and pricing

  • Free tools need no account.
  • Mastering: 3 per month per IP without an account, 1 per day with a free account.
  • Register an agent once (POST /api/agents/register) and get 3 trial masters; afterwards 1 token = 1 full PRO master. Tokens never expire. Live prices: GET /api/tokens/packages.
  • Rules for agents, error codes and the payment protocol: https://magicmaster.pro/agents.md.

Links

Terms https://magicmaster.pro/terms · Privacy https://magicmaster.pro/privacy · Public OpenAPI https://magicmaster.pro/openapi-public.json · Support support@magicmaster.pro


По-русски

Удалённый MCP-сервер Magic Master: мастеринг под целевую громкость (LUFS, True Peak), снятие цифрового следа Suno/Udio, паспорт результата прямо в чате. Ставить ничего не нужно — адрес https://magicmaster.pro/mcp. Подробности и подключение по-русски: https://magicmaster.pro/mcp.

Tools it offers (11)

What this server listed when ahel dialed its public endpoint in Sep 2026, with no key and no account of yours. The names are the server’s own.

  • get_service_info
  • list_presets
  • get_pricing
  • check_limits
  • analyze_track
  • clean_ai_trace
  • preview_styles
  • master_track
  • batch_master
  • get_job
  • create_topup_link

Signals

Last commit
Sep 2026
Weekly downloads
139
Advanced
Delivery
mastering MCP server → your ahel gateway (mcp.ahel.ai) → every connected AI client.
Catalog kind
mcp-server
Gateway key
pro-magicmaster-mastering
Source
github.com/denisok-ai/magicmaster-mcp
Hosted endpoint
https://magicmaster.pro/mcp