Sonicaura processes audio using GPU-accelerated AI. It exposes atomic, 1-to-1 semantic tools optimized for LLM reasoning.
- Stem Separation — split a track into 2, 4, or 6 stems
- Quick Splitter — isolate any single instrument (guitar, piano, bass, drums)
- Chord Recognition — detect chord progressions from the track
- MIDI Export — convert isolated instruments into editable MIDI files
- Denoise — remove noise from music or aggressively clean up speech
- Reverb — add spatial reverb with presets: vocal_booth, studio_room, bathroom, concert_hall, church
- URL:
https://api.sonicaurastudio.com/mcp - Transport: http (Streamable HTTP)
- Auth: OAuth 2.0
| Tool | Description |
|---|---|
get_upload_url |
Get a presigned URL for large file uploads (bypasses LLM limits). |
extract_vocals |
2 stems: isolate Vocals and Instrumental (Karaoke). |
split_4stems |
4 stems: Vocals, Drums, Bass, Other. |
split_6stems |
6 stems: Vocals, Guitar, Piano, Bass, Drums, Other. |
extract_single_instrument |
Quick Splitter: isolate a specific instrument. |
extract_chords |
AI Chord Recognition: detect chord progression. |
export_midi |
AI MIDI Export: convert an isolated instrument into MIDI. |
denoise_audio |
Remove noise (use mode="music" or mode="speech"). |
apply_reverb |
Add spatial reverb. |
check_task_status |
Poll a task (requires only task_id). Auto-detects all outputs. |
list_audio_tools |
List available capabilities. |
get_my_credits |
Check remaining credits. |
check_service_status |
Check server health. |
Add the server to your MCP client config (Cursor, Claude Desktop, VS Code, etc.):
{ "mcpServers": { "sonicaura": { "type": "http", "url": "https://api.sonicaurastudio.com/mcp" } } }
Then ask your AI client things like:
- "Extract the vocals from this song"
- "Split this track into 6 stems"
- "Isolate the guitar track and export it as a MIDI file"
- "Detect the chords in this song"
- "Denoise this music but keep the background music"
- "Add studio reverb to this vocal"
Prefer a visual interface? Use the same engine from the browser:
Processing uses credits. Get an account and credits at sonicaurastudio.com
MIT