@ericdisero/aurora-mcp-server
v0.3.0
Published
MCP server for Aurora, the AI audio workbench. Your AI agent (Claude, Cursor, or any MCP client) drives music generation, covers, sample manufacturing, 7-stem separation, and DAW-ready stem organization. Files land on disk.
Maintainers
Readme
@ericdisero/aurora-mcp-server
Stdio MCP server for Aurora, the AI audio workbench. Gives Claude (or any MCP client) full control of a local music library: generate tracks, style-transform covers, manufacture key/tempo-requested samples, split anything into 7 stems. 31 tools. Files land on disk in real project folders the Aurora app shows live.
Install
{
"mcpServers": {
"aurora": {
"command": "npx",
"args": ["-y", "@ericdisero/aurora-mcp-server"],
"env": {
"SUNO_API_KEY": "<your sunoapi.org key>",
"MVSEP_API_KEY": "<your mvsep key>"
}
}
}
}Claude Code: claude mcp add aurora -- npx -y @ericdisero/aurora-mcp-server
The env block is optional if keys are stored via the CLI (aurora keys --suno-api-key … --mvsep-api-key … → ~/.aurora/config.json).
Highlights
- Standalone — works against Aurora's database + project folders directly; the desktop app doesn't need to be running.
- Background jobs —
background: trueon generate/cover/sounds/split returns a jobId;aurora_get_job_statuspolls, downloads, and registers results. Jobs survive restarts. - Streaming preview — in-flight generations expose
streamUrls: listenable ~30-45s in, minutes before files land. - Progressive stems — splits land vocals/drums/bass stems as each separation job finishes; everything-else last.
- Cost discipline built in —
aurora_get_creditsis free; destructive ops requireconfirm: true; re-splitting an already-split asset is refused.
Full docs: github.com/EricDisero/aurora-mcp. CLI sibling: @ericdisero/aurora-cli.
MIT. Copyright Blueprint Online Learning Inc.
