# Lyria 3.5 (music)

> Lyria 3.5 writes a full song from one text prompt. The prompt carries the style, the vocals, the lyrics and the section timestamps.

Status: roasOS API v1.1. Catalog `model-catalog@2026-10-06.v5`, pricing `customer-pricing@2026-11-05.v15`. The endpoint `POST /v1/music/lyria-3.5` is live. `POST /v1/generations` has no `music` type, so this endpoint is the only path to this model.

## Endpoint
POST https://api.roasos.com/v1/music/lyria-3.5
Auth: Authorization: Bearer roasos_sk_… (scope api:write). Idempotency-Key required. quote_only / quote_id / webhook_url as in https://docs.roasos.com/llms.txt.

## Good at
Make a song ad from one prompt. Write the genre, tempo, instruments, vocal style and the words to sing; the model writes and performs the song. Songs can be up to a few minutes long. The model sings in English, German, Spanish, French, Hindi, Japanese, Korean and Portuguese.

The price is $0.10 per track, flat. There is no duration field, so the quote is exact and nothing reconciles. The endpoint states no fixed output length, so roasOS publishes none: read the length of the file the job returns.

There is no separate lyrics field. Put the lyrics in `prompt`. The output can include a `lyrics` string, which is the words the model sang.

roasOS states no rights position on the music this model generates. Read the terms of the model and of fal before you publish a track.

## Pricing
Unit: track. Pricing key: `lyria-3.5`.

| request | credits |
|---|---|
| `per_track` | 2 |

Credits come from the roasOS pricing table (`MODEL_COSTS`) through the catalog. No other field changes the price.

## Modes
- `text-to-music`
Provider endpoints: `google/lyria-3.5`.

## Request fields
| field | type | required | default | rule |
|---|---|---|---|---|
| `image_url` | string | no | — | https URL; ≤ 4,096 characters |
| `prompt` | string | yes | — | ≥ 1 character; ≤ 5,000 characters |
| `quote_id` | string | no | — | UUID |
| `quote_only` | boolean | no | `false` | — |
| `wait_timeout_seconds` | integer | no | `0` | 0–60 |
| `webhook_url` | string | no | — | https URL; ≤ 4,096 characters |

The schema is strict: a field that is not in this table returns `400 unsupported_option`. Field names are the roasOS role names, one name per role across every family.

## Media
- `image_url` — first frame (or the single source image)

| role | items | max size | content types | duration |
|---|---|---|---|---|
| `image_url` | ≤ 1 | 10 MB | `image/jpeg`, `image/png`, `image/webp` | — |

- Refused: `avatar_id`, `elements`, `product_id`, `product_images`, `references`, `reference_images`, `end_image_url`, `scene_id`, `scene_image`, `video_urls`, `video_url`, `audio_urls`, `audio_url`, `voice_id`. A refused role returns `400 unsupported_option`.
- Every URL must be an immutable object on `https://cdn.roasos.com/…` that your workspace owns; any other URL is `400 reference_media_not_workspace_owned`. Upload with `POST /v1/uploads` (direct POST of the bytes to `put_url`, or a `source_url` import), then use the returned CDN URL.

## Prompt
Write the brief, not the adjective. Name the genre, the BPM, the instruments, the vocal style and the hook. To control the length and the structure, write a length ("a 30-second song") or section timestamps:

"[0:00-0:10] Intro: soft piano. [0:10-0:25] Verse: 'Morning light on the kitchen floor'. [0:25-0:30] Chorus: 'Rise up and shine today'."

Align the timestamps to the cuts of the ad when the song carries the whole soundtrack.

`image_url` is mood inspiration, not a first frame: the model reads the feel of the picture, not its contents.

## Not exposed
- `negative_prompt`: pinned to `provider-deprecated: fal states negative prompting is not supported`.

## Result
Result kind `audio`. Always present: `url`. Present when the run produces it: `timestamps`.
```json
{ "data": { "id": "gen_…", "kind": "audio", "model": "lyria-3.5", "status": "completed", "result": {"timestamps":[{}],"url":"https://cdn.roasos.com/workspaces/ws_01/result"} } }
```
Files live on `https://cdn.roasos.com` and are immutable; a result URL can be a media input to the next call. Poll `GET /v1/jobs/{id}/status` → `{ id, status, updated_at }`, then `GET /v1/jobs/{id}/result` (`409 run_not_completed` until the job is `completed`, `failed` or `canceled`). Statuses: `queued → processing → completed | failed | canceled`.

## Example
```bash
# 1. quote — nothing is spent
curl -X POST https://api.roasos.com/v1/music/lyria-3.5 \
  -H "Authorization: Bearer roasos_sk_live_…" \
  -H "Idempotency-Key: 0f1c9e2a-lyria-3.5-0001" \
  -H "Content-Type: application/json" \
  -d '{"quote_only":true,"prompt":"A woman holds the bottle at eye level in a bright kitchen, morning light, static camera."}'
# → 200 { "data": { "kind": "quote", "quote_id": "q_7Hk…", "credits": 2, "expires_at": "…", "confirmation": { "model": "lyria-3.5", "example": "per_track" } } }

# 2. run — the same body and the same Idempotency-Key, quote_id instead of quote_only
curl -X POST https://api.roasos.com/v1/music/lyria-3.5 \
  -H "Authorization: Bearer roasos_sk_live_…" \
  -H "Idempotency-Key: 0f1c9e2a-lyria-3.5-0001" \
  -H "Content-Type: application/json" \
  -d '{"quote_id":"q_7Hk…","prompt":"A woman holds the bottle at eye level in a bright kitchen, morning light, static camera."}'
# → 201 { "data": { "id": "gen_…", "kind": "audio", "model": "lyria-3.5", "status": "queued", "credits_reserved": 2, "links": { "status": "/v1/jobs/gen_…/status", "result": "/v1/jobs/gen_…/result" } } }

# 3. poll
curl https://api.roasos.com/v1/jobs/gen_…/status -H "Authorization: Bearer roasos_sk_live_…"
curl https://api.roasos.com/v1/jobs/gen_…/result -H "Authorization: Bearer roasos_sk_live_…"
```

## Today
The endpoint is `POST /v1/music/lyria-3.5`.

It is live.

## See also
https://docs.roasos.com/models/music/lyria-3.5 · https://api.roasos.com/v1/catalog/music/lyria-3.5 · google/lyria-3.5