# The AI Music API for Songs, Jingles and Soundtracks

> Send Revana an idea, a style or your own lyrics, and Suno's newest model writes and sings a full song or plays an instrumental. Every job returns two takes, so you can keep the better one. A job costs $0.08 with [Suno V6](https://revana.si/models/suno-v6) (prices as of November 2026), a job that fails costs nothing, and the request shape matches every other model on your key. Prices as of 2026-11-21.

- Page: https://revana.si/music-api

## Request

```bash
# One Idempotency-Key per job, made once: resend with the same $KEY after a timeout, a 429 or a 5xx
KEY=$(uuidgen)
curl https://api.revana.si/v1/generations \
  -H "Authorization: Bearer $REVANA_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $KEY" \
  -d '{
  "model": "suno-v6",
  "input": {
    "prompt": "An upbeat song for a coffee shop opening on a sunny Saturday morning",
    "style": "indie pop, warm female vocals, acoustic guitar, hand claps"
  },
  "max_cost_usd": 0.08
}'
```

## Inputs

- `prompt`: the idea, up to 3,000 characters
- `style`: genre, instruments, voice
- `lyrics`: words to sing ([Verse], [Chorus] tags)
- `instrumental`: true for no vocals (never with lyrics)
- Every request needs a style or lyrics, or both

## Batches (USD, November 2026)

| Jobs | Tracks | Total |
|---|---|---|
| 1 | 2 | $0.08 |
| 10 | 20 | $0.80 |
| 100 | 200 | $8.00 |
| 1000 | 2000 | $80.00 |

## Python

```python
import os, time, uuid, requests

API = "https://api.revana.si/v1"
H = {"Authorization": f"Bearer {os.environ['REVANA_API_KEY']}"}

# made once: resend with the same key after a timeout and nothing is made or charged twice
key = str(uuid.uuid4())
job = requests.post(f"{API}/generations", headers={**H, "Idempotency-Key": key}, json={
    "model": "suno-v6",
    "input": {"prompt": "A short jingle for a bike shop", "style": "pop punk, bright guitars"},
    "max_cost_usd": 0.08,
}).json()

while job["status"] not in ("succeeded", "failed"):
    time.sleep(5)
    job = requests.get(f"{API}/generations/{job['id']}", headers=H).json()

if job["status"] == "succeeded":
    for n, path in ((1, "content"), (2, "content/2")):  # two takes per job
        song = requests.get(f"{API}/generations/{job['id']}/{path}", headers=H)
        open(f"take-{n}.mp3", "wb").write(song.content)
    print("charged", job["cost_usd"])
```

## FAQ

### What is an AI music API?

An AI music API turns a text description, a style or a set of lyrics into a finished track your app can download. On Revana you call Suno's newest model with one key and the same request shape you use for video and images.

### How much does a generated song cost?

One job costs $0.08 on Revana and returns two tracks, so each song works out at 4 cents (prices as of November 2026). You pay from prepaid credit with no subscription, and a failed job is not charged.

### Why do I get two tracks?

Every music job returns two takes of the same request, so you can pick the stronger one or offer both to your users. Download the first from /content and the second from /content/2 once the job has succeeded.

### Can I use my own lyrics?

Yes. Send your words in lyrics along with a style for the sound, and put section tags such as [Verse] and [Chorus] on their own lines to shape the song. The model writes the melody and sings your words.

### Can it make music without vocals?

Yes. Set instrumental to true and describe the music in prompt and style, such as soft piano with vinyl crackle. Do not send lyrics in the same request: an instrumental request with lyrics is refused before anything is charged.

### Can I use the songs commercially?

Commercial use depends on the music model maker's own terms, so read them before you publish or sell a track. The model page names its maker.

### Am I charged if a song fails?

No. The price is held when you submit and released if the job fails, so a failed job shows as $0.00. Resend it with a new Idempotency-Key and you are charged once, only when the tracks are delivered.

