Suno music API
Generate songs and lyrics, then extend, cover, separate or export your own tracks. Availability and current prices come from the model catalog.
Operation
Use an Audio API key with model=suno visible in GET /v1/models. Every POST returns a gateway task_id. Poll GET /v1/music/tasks/{task_id} or GET /v1/task/{task_id} until success; HTTP200 alone is not completion.
| Operation | POST | Parameters |
|---|---|---|
| Generate music | /v1/music/generations | version=v6 | v6-mini | v6-wild; custom; prompt; style; instrumental; duration; max_mode |
| Generate lyrics | /v1/music/generations/lyrics | prompt; lyrics_model=classic | remi |
| Extend | /v1/music/generations/extend | task_id; audio_index; continue_at; version; custom; prompt; tags; duration_s |
| Cover | /v1/music/generations/coverSong | task_id; audio_index; version; custom; prompt; tags; duration_s |
| Separate vocals | /v1/music/generations/stems | task_id; audio_index; stem_type=lead_vocal; audio_format |
| Export MP3 / WAV | /v1/music/generations/download | task_id; audio_index; formats=[mp3,wav] |
Custom lyrics
custom=false: prompt describes the song. custom=true: prompt contains lyrics; style describes the music. Duration10–360 seconds and max_mode require custom=true. Max is2× the call price. Standard generation, extension, cover, lyrics, stems and export have separate prices.
Per accepted call; returned track count does not multiply the price
For vocal songs, start with 60 seconds. Short requests may not contain all lyrics.
curl https://apixly.ai/v1/music/generations \
-H "Authorization: Bearer YOUR_AUDIO_API_KEY" \
-H "Idempotency-Key: YOUR_STABLE_REQUEST_UUID" \
-H "Content-Type: application/json" \
-d '{"model":"suno", "version":"v6", "custom": true, "instrumental":false,
"prompt":"[Verse]\nPaper boats along the stream", "style":"acoustic pop, piano",
"duration":60, "audio_format":"mp3", "max_mode":false}'Source track
task_id must be a completed song's owned gateway UUID, never a provider task ID. audio_index is the original1-based index in output.tracks. The original provider account remains pinned. continue_at must be at least1 and below the source duration. Export/stem/lyrics results cannot become source songs.
curl https://apixly.ai/v1/music/generations/extend \
-H "Authorization: Bearer YOUR_AUDIO_API_KEY" \
-H "Idempotency-Key: ANOTHER_STABLE_REQUEST_UUID" \
-H "Content-Type: application/json" \
-d '{"model":"suno", "task_id":"OWNED_GATEWAY_TASK_UUID", "audio_index":1,
"version":"v6", "custom":true, "prompt":"[Chorus]\nWe keep moving gently on",
"tags":"acoustic pop", "continue_at":30, "duration_s":60, "audio_format":"mp3"}'Result
A successful song returns every file under output.result[].audio_urls, with source indices in output.tracks. Lyrics return output.text_result without an audio file. In Studio, completed files are archived to your private cloud library. Keep Idempotency-Key unchanged when checking an ambiguous POST; do not generate another task to recover a result.
curl https://apixly.ai/v1/music/tasks/OWNED_GATEWAY_TASK_UUID \
-H "Authorization: Bearer YOUR_AUDIO_API_KEY"
# Poll until status=success. Lyrics: output.text_result.
# Songs: output.result[].audio_urls; source indices: output.tracks[].audio_index.
# Unknown: POST /v1/task/OWNED_GATEWAY_TASK_UUID/reconcile; never repeat generation.Accepted tasks cannot be cancelled. Unknown/review/storage_required means check or recover the original task; it does not authorize a new paid submission.
Format
- MP3 / WAV: up to 360 seconds; 80 MiB per file, 320 MiB per operation.
- MP3 and WAV, up to 360 seconds. Files are bounded to 80 MiB each and 320 MiB per operation. Formats in one export are billed as one operation.