Short answer: call
POST /api/music/separatewith the sourcesong_idandstem_mode: "twelve". Poll the returned task and save every completed stem ID. Each stem can be downloaded as audio or exported as MIDI independently.
Two stems or twelve?
| Mode | Output | Price | Best use |
|---|---|---|---|
two | Vocal + instrumental | 0.6 CNY | Covers, karaoke, acapella |
twelve | Drums, bass, guitar, keys and more | 3 CNY | Mixing, re-arrangement, MIDI |
Submit twelve-stem separation
curl -X POST "https://www.suno-api.io/api/music/separate" \
-H "Authorization: Bearer $SUNO_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"song_id": "YOUR_SOURCE_SONG_ID",
"stem_mode": "twelve",
"model": "suno-v6"
}'Do not submit the same source repeatedly while the existing separation is rendering.
Save stems by role
Store each returned stem ID with a stable role such as drums, bass, guitar or vocals. Do not depend on display order.
Download audio or MIDI
# Audio URL
curl -X POST "https://www.suno-api.io/api/music/download-url" \
-H "Authorization: Bearer $SUNO_API_KEY" \
-H "Content-Type: application/json" \
-d '{"song_id":"stem_clip_id","kind":"song"}'
# MIDI file
curl -X POST "https://www.suno-api.io/api/music/download-file" \
-H "Authorization: Bearer $SUNO_API_KEY" \
-H "Content-Type: application/json" \
-d '{"song_id":"stem_clip_id","kind":"midi"}' \
--output stem.midFAQ
When should I use twelve stems instead of two?
Use twelve only when individual instruments need separate processing. Two stems are cheaper for ordinary covers and karaoke.
Can each separated stem export MIDI?
Yes, after that stem is complete. Use its clip ID, not the original song ID.
Why can one stem sound empty?
The source may not contain that instrument, or it may have no clear pitched notes.
Read the separation API, MIDI guide, and mixing guide.
¥0.6 per generation, 2 tracks, unlimited downloads
Runs on the Suno V6 engine. Download MP3 or WAV as often as you like. Signing up needs an email address, not a phone number.


