Overview

The Audius Atris MCP Server implements a Model Context Protocol (MCP) bridge for interacting with the Audius Atris network through language models such as Claude. It exposes a small API surface that makes it possible to stream tracks, send tips to artists, and drive interaction workflows from an LLM agent or any client able to speak MCP. This is useful for building conversational music assistants, automated promotion agents, and fan-engagement bots that integrate audio streaming and crypto-enabled tipping with natural-language tools.

The server acts as the glue between an MCP-capable model and Audius/Atris primitives: playback links, metadata lookups, and on-chain or off-chain tipping flows. It provides developer-friendly endpoints, common wiring for environment configuration, and an HTTP surface that the model can call as “tools” during a conversation. This lets teams prototype and deploy music-focused agents quickly without building low-level Audius integration logic from scratch.

Features

Exposes MCP-style tools for streaming, tipping, and fan interactions
Converts model tool calls into Audius/Atris actions (metadata, stream URLs)
Example endpoints for streaming a track, tipping an artist, and querying metadata
Configuration via environment variables, with Docker support for local testing
Minimal API surface designed to be called by LLMs or web clients
Logging and basic error handling to aid debugging during development

Installation / Configuration

Clone the repository and install dependencies. The project includes a straightforward startup path using Node.js (or you can run it with Docker if provided).

Clone and run locally (Node.js example):

git clone https://github.com/glassBead-tc/audius-mcp-atris.git
cd audius-mcp-atris

# If the project uses Node.js
npm install
npm run start

Or using Docker (if the repo supplies a Dockerfile/docker-compose):

git clone https://github.com/glassBead-tc/audius-mcp-atris.git
cd audius-mcp-atris

# build and run with Docker
docker build -t audius-mcp-atris .
docker run -p 3000:3000 --env-file .env audius-mcp-atris

Example .env template (create .env in the project root):

# Server
PORT=3000

# Audius / Atris endpoints
AUDIOUS_API_URL=https://api.audius.co
ATRIS_NODE_URL=https://atris.example.com

# Keys (use secure secret management in production)
MCP_API_KEY=your-mcp-api-key
ATRIS_PRIVATE_KEY=0x...

# Optional: database or caching
DATABASE_URL=sqlite://./data.db

After configuration, start the server and check the health endpoint:

curl http://localhost:3000/health

Available Tools / Resources

The MCP server exposes a small set of endpoint “tools” intended to be called from an LLM or a client. Typical endpoints include:

GET /tracks/:id — retrieve metadata and stream URL for a track
POST /stream — request a streaming session or playback URL for a track
POST /tip — initiate a tip (micro-transaction) to an artist
POST /interact — generic interaction endpoint for chat or fan engagement workflows
GET /health — health check for the server

Example cURL: request a stream URL

curl -X POST http://localhost:3000/stream \
  -H "Authorization: Bearer $MCP_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"trackId":"12345","quality":"high"}'

Example cURL: tip an artist

curl -X POST http://localhost:3000/tip \
  -H "Authorization: Bearer $MCP_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"artistId":"artist_abc","amount":"0.01","currency":"ETH","memo":"Love this track!"}'

Note: Exact request/response formats will match the server’s API schema in the repository; adapt the payloads accordingly.

Environment Variables Reference

Variable	Purpose
PORT	HTTP port the server listens on
AUDIOUS_API_URL	Base URL for Audius API or metadata provider
ATRIS_NODE_URL	Endpoint for Atris node or provider
MCP_API_KEY	API key to authenticate MCP/tool calls
ATRIS_PRIVATE_KEY	Private key used for signing tip transactions (use secure vault in prod)
DATABASE_URL	(Optional) DB connection for persistence or caching

Use Cases

Conversational Streaming Bot: Connect a Claude-style assistant to the MCP server so fans can ask for a specific track by name. The model queries /tracks, receives metadata and a streaming URL, and responds with playback options—enabling a chat-driven listening experience.
Native Tip Flow: Implement a “tip the artist” tool in your LLM agent. When a fan types “I want to tip the producer 1 USD,” the model calls /tip with the appropriate artist ID and amount. The server signs and executes the payment flow (on-chain or via Atris), returning a transaction receipt the model can share.
Release Promotion & Fan Engagement: Automate release announcements in chat. The model can call /interact to enroll fans in a drip campaign, fetch analytics for a track, or send preview stream links to engaged users.
Moderation & Curation: Use the /tracks endpoint to fetch metadata and verify content before allowing automated promotions. Combine with simple blacklist/whitelist logic in the server for safe content routing.

Tips for Developers

Run locally first with the included .env and a test account on Audius/Atris to avoid expenditure of real funds.
Use secure secret storage for private keys and production API keys.
Log tool calls and model decisions during early testing so you can iterate on prompts and endpoint behavior.

For implementation details and the exact API schema, see the repository at https://github.com/glassBead-tc/audius-mcp-atris.

Audius Atris MCP Server: Interact, Stream, Tip