One song, then a handle
Write genre, instruments, and mood. Lyrics, title, BPM, and length are optional. Length is a 1-240 second hint, not a hard cap and not a per-second bill.
Generate a full song with Google Lyria (Flow Music), then extend, replace, cover, split stems, or export wav. Playground and API on AIReiter.
| Specification | Credits | Price (USD) |
|---|---|---|
| Model & ModalityFlow Music | 7.2per generation | $0.072per generation |
| Model & ModalityFlow Music Lyrics | 2.4per generation | $0.024per generation |
| Model & ModalityFlow Music Extend | 7.2per generation | $0.072per generation |
| Model & ModalityFlow Music Replace | 7.2per generation | $0.072per generation |
| Model & ModalityFlow Music Cover | 7.2per generation | $0.072per generation |
| Model & ModalityFlow Music Stems | 7.2per generation | $0.072per generation |
| Model & ModalityFlow Music Upload | 1.2per generation | $0.012per generation |
| Model & ModalityFlow Music Download | 2.4per generation | $0.024per generation |
| Model & ModalityFlow Music Video | 2.4per generation | $0.024per generation |
Prices follow the model configuration. Reference images are not billed unless the model says so.
Flow Music is Lyria on AIReiter. Generate one full song from a style prompt, keep the clip_id, then extend, replace a range, cover, split stems, export, or render a video without starting over.
Write genre, instruments, and mood. Lyrics, title, BPM, and length are optional. Length is a 1-240 second hint, not a hard cap and not a per-second bill.
Extend from a timestamp. Replace only the bars that fail. Cover restyles the whole clip; strength (0-1, default 0.5) is how far the new version may drift.
Stems returns a zip for a DAW. Download exports the same clip as mp3 or wav. Video renders an mp4 from simple, modern, or player templates.
If you only have mp3/wav/m4a, run Upload first (20 MB, 480 seconds). Every later type on this page needs that clip_id.
From the public API docs on this site. Prompt: Short acoustic guitar loop, no vocals. Version: Lyria 3 Pro. Length hint: 60 seconds. Audio plays in the playground above; the image is the cover from that task. We do not freeze a credit number here - use the live table.
Listen in the playground
Choose Lyria when the first generate is a bed you will edit. Choose Suno when you want a sung identity and two candidates from one prompt.
Describe the sound, optionally paste lyrics, generate once. Keep clip_id even if you only need the preview today.
Replace a start-end range when the chorus works and the bridge does not. Extend when you need more seconds on the same clip.
Stems is a zip, not a playable mix. Download wav when the streaming preview is not the delivery file.
Video turns the same clip_id into an mp4 template. It is not a full music-video model - it is a shareable frame around the song.
Tabs map to flow_music* model IDs. Follow-up types need clip_id from generate or upload.
Generate is the song. Lyrics is a title and verse/chorus sheet with no audio. Lyria 3 Pro is default; Lyria 3.5 is the newer vocal model at the same listed price.
All three need clip_id and an instruction. Extend: from + duration. Replace: start + end seconds. Cover: strength 0-1.
Stems = zip. Download = mp3 or wav. Video = mp4 preset. None of these invent a new composition; they package the clip you already have.
Only when you have a file and no clip_id. After upload, the other edit types become available.
Credit amounts live in the pricing table above this section. Length is a hint, not a billed duration.
| Typical Lyria access | Flow Music on this page | |
|---|---|---|
| What you get back | A generated clip | A clip plus clip_id used for every later step |
| After the first song | Often a new generate | Extend, replace a range, cover, stems zip, wav/mp3 export, or an mp4 template |
| How you pay | Cloud or studio quotas | Credits per successful task, listed in the table above |
| API | Provider-specific endpoints | POST /api/openapi/submit with flow_music* IDs, then POST /api/openapi/query |
| UI | Docs or a console | Playground on this page with the same fields |
Parameter tables live in the API reference. Billing is per successful task, not per second of length.
POST /api/openapi/submit with model flow_music or flow_music_lyrics / extend / replace / cover / stems / upload / download / video.
POST /api/openapi/query with your out_task_id. A generate output item includes audio, cover, title, text, and clip_id.
Pass it on every edit/export call. Upload is the only way to mint one from a file you already have.
Bearer key from /keys. Music shares the image and video credit balance.
Suno if you want a sung identity. Tools page if you want every audio model in one UI.
Prompt-to-song with lyrics, extend, and cover from an audio file. Usually two candidates per generate.
02AI Audio toolsWorking UI for the full audio catalog.
03Credit pricingSame credit packs as the rest of the site.
04Flow Music API referenceSubmit example for flow_music, including version and length notes.
| Control | Type | Notes | |
|---|---|---|---|
| 01 | version | string | — |
| 02 | prompt | text | Describe the sound. Add lyrics below if you want singing. |
| 03 | lyrics | text | — |
| 04 | title | text | — |
| 05 | bpm | text | — |
| 06 | length | integer | — |
Style prompt first. Length is a hint. Submit and wait for clip_id.
Extend, replace, or cover with an instruction. Set timestamps only for the part that should change.
Stems zip, mp3/wav, or mp4 from the same clip_id.
flow_music, flow_music_lyrics, flow_music_extend, flow_music_replace, flow_music_cover, flow_music_stems, flow_music_upload, flow_music_download, flow_music_video
POST to /api/openapi/submit with the model, a unique out_task_id, and parameters.
POST to /api/openapi/query with the same out_task_id until the task completes.
Save outputs in your own storage rather than hot-linking them.
Flow Music is the Lyria text-to-music family on AIReiter. Generate uses flow_music. Follow-up types use flow_music_* IDs on this page.
Generate and upload return it. Extend, replace, cover, stems, download, and video all require it. If you only have a file, run Upload first.
No. Length is a 1-240 second hint. Upstream may not deliver that exact duration. You are billed per successful task, not per second.
A zip of separated tracks, not a playable mix. Use Download for mp3 or wav of the full clip.
Lyria 3 Pro is the default. Lyria 3.5 is the newer vocal model. They share the generate price in the table above.
POST /api/openapi/submit with the flow_music* model ID, then POST /api/openapi/query with out_task_id. Examples: docs.aireiter.com/en/api-reference/audios/flow-music/generation