🎤

AI Song Generation

Lyrics plus a style prompt in, a finished song out.

Write the lyrics, describe the genre and mood, choose 10-120 seconds. Get an MP3 with vocals, or set instrumental=true for backing music. Powered by ACE-Step 1.5 (MIT licence).

🚀 Open Song Generator API reference

What it does

Songs with sung vocals or instrumental tracks from one endpoint.

Vocals from your lyrics

Sung in English, Hindi, Chinese, Japanese or Korean. Use [verse] / [chorus] tags to shape the structure.

Instrumental mode

Set instrumental to true (lyrics optional) for background music for reels, ads and videos.

10-120 seconds

Pick the length per request. Output is an MP3.

Your rights

Only submit lyrics you have the right to use.

Listen

Sample outputs from the ACE-Step song engine.

🎧
Upbeat indie rock · 15 s sample
🎧
Soft acoustic ballad · 15 s sample

API

Authenticate with Authorization: Bearer YOUR_API_KEY (create a key in the dashboard). Short songs usually return output_url in the response. Longer ones return status: "processing"; poll GET /v1/song/{generation_id}.

POST /v1/song/generate

JSON body. Returns the MP3 output_url inline, or a generation_id to poll.

FieldTypeRequiredDescription
promptstringyesStyle description: genre, mood, instruments (max 500 characters).
lyricsstringno*Lyrics to sing (max 2,000 characters). *Required unless instrumental is true.
durationintegernoLength in seconds, 10-120 (default 30).
languagestringnoVocal language: en (default), hi, zh, ja, ko.
instrumentalbooleannoSet true for a track with no vocals.
curl -X POST https://api.pixelapi.dev/v1/song/generate \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{"prompt": "upbeat indie pop, electric guitar, drums", "lyrics": "[verse]\nCity lights are calling me\n[chorus]\nWe run all night", "duration": 60, "language": "en", "instrumental": false}' # quick jobs return output_url inline; otherwise status is "processing": curl https://api.pixelapi.dev/v1/song/GENERATION_ID -H "Authorization: Bearer YOUR_API_KEY"

POST /v1/audio/generate (instrumental shortcut)

Same engine and price with instrumental forced on.

FieldTypeRequiredDescription
promptstringyesDescribe the music (genre, mood, instruments).
durationintegernoSeconds, 10-120 (default 30).
curl -X POST https://api.pixelapi.dev/v1/audio/generate \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{"prompt": "calm corporate background music, soft piano and strings", "duration": 30}' # if the response has status "processing", poll: curl https://api.pixelapi.dev/v1/song/GENERATION_ID -H "Authorization: Bearer YOUR_API_KEY"

Pricing

Billed in credits. 1 credit = $0.001 on the Starter plan; the cash value of a credit depends on your plan. Failed jobs are refunded.

Per song or track
28 credits
≈ $0.028 on Starter. Any length from 10 to 120 seconds, with or without vocals.

FAQ

How much does a song cost?

28 credits per song or instrumental track, for any length from 10 to 120 seconds. 1 credit = $0.001 on the Starter plan.

Which model generates the songs?

ACE-Step 1.5, released under the MIT licence.

Can I make music without vocals?

Yes. Set instrumental to true on /v1/song/generate, or call /v1/audio/generate.

Which vocal languages are supported?

English, Hindi, Chinese, Japanese and Korean.

What format is the output?

MP3.

Ready to try AI Song Generation?

New accounts get 300 free trial credits (one trial per person), no card required.

🚀 Open Song Generator ← All tools