Udio API: Generate AI Music with Udio Programmatically

Udio is the second AI music model people ask about after Suno. The catch: Udio Inc. doesn't ship a public API. No self-service keys, no SDK, no docs.

This post covers how to access Udio programmatically through sunor, the differences between Udio and Suno output, and when each model is the better choice.

Does Udio Have an API?

No. Udio Inc.'s own help center states it directly: "We know there's keen interest, but we don't currently offer a public API." Music generation is only available through the Udio web app — there's no developer-facing endpoint from Udio Inc.

sunor now supports Udio alongside Suno. Same REST API, same auth, same credits system. You set model: "udio" instead of model: "suno" and everything else works the same way.

import requests, time
 
response = requests.post(
    "https://sunor.cc/api/v1/task",
    headers={"x-api-key": "sk_live_YOUR_API_KEY"},
    json={
        "model": "udio",
        "task_type": "music",
        "input": {
            "prompt": "lofi hip hop, chill, rainy night, jazz piano",
        },
    },
)
task_id = response.json()["data"]["task_id"]
 
while True:
    result = requests.get(
        f"https://sunor.cc/api/v1/task/{task_id}",
        headers={"x-api-key": "sk_live_YOUR_API_KEY"},
    )
    data = result.json()["data"]
    if data["status"] in ("success", "failure"):
        break
    time.sleep(5)
 
if data["status"] == "success":
    for song in data["output"]["result"]:
        print(song["title"], "→", song["song_path"])

Two endpoints. A short style description in, audio files out. One Udio call returns about two generated songs — pick whichever you like best, both billed as one task.

What Makes Udio Different

Udio leans into production polish and natural vocal delivery on shorter prompts. Where Suno is the more versatile generalist (verse-by-verse lyrics control, multiple task types, voice cloning), Udio's strength is sounding finished out of the box.

Three quick characteristics worth knowing:

Short prompts win. Udio has a GPT prompt-expansion layer that turns 5–15 word style descriptions into full prompts. Long, detailed prompts can fight that expansion. "lofi hip hop, chill, rainy night, jazz piano" is the sweet spot — give it a vibe, let the model fill in the structure.

Vocals sound polished. On indie folk, dream pop, downtempo electronica, and similar genres, Udio's vocal delivery often sounds more "produced" than Suno's out of the box.

Two tracks per call. One Udio task returns about two completed songs. Both are billed as one call — useful for picking the take you like rather than re-generating from scratch.

Generation Modes

Three modes via the lyrics_type field:

// 1. Default — Udio writes lyrics from your style prompt
{
  "model": "udio",
  "task_type": "music",
  "input": {
    "prompt": "indie folk, acoustic, melancholy"
  }
}

// 2. User-supplied lyrics — pass your own lyrics verbatim
{
  "model": "udio",
  "task_type": "music",
  "input": {
    "prompt": "indie folk, acoustic, melancholy",
    "lyrics": "[Verse]\nWalking down the midnight road\n[Chorus]\nWe are the dreamers in the night",
    "lyrics_type": "user"
  }
}

// 3. Instrumental — no vocals
{
  "model": "udio",
  "task_type": "music",
  "input": {
    "prompt": "cinematic orchestral, sweeping strings, dramatic",
    "lyrics_type": "instrumental"
  }
}

Udio vs Suno

If you already have a Suno API integration through sunor, adding Udio is a one-line change — same auth, same task lifecycle, same output format on the customer side.

When to Pick Which

Use Suno when you need:

• Verse-by-verse lyrics control with explicit section markers
• Audio upload (cover / extend an existing track)
• Concatenating clips into a single longer track
• Lyrics-only generation as a separate task
• The widest range of genres in one model

Use Udio when you need:

• Polished production aesthetic on shorter prompts
• Natural vocal delivery on indie / folk / dream pop / downtempo
• Lower cost per call (half of Suno music)
• Quick "generate a vibe" workflows where the GPT expansion does the heavy lifting

For most apps, the right answer is "support both, let the user pick." That's why we built sunor as a unified API rather than two parallel ones.

Pricing

Pay-as-you-go credits. Same system as Suno on sunor. No subscription, no minimum spend.

Failed tasks are automatically refunded. Full pricing on the Pricing page.

Get Started

Sign up gets you 25 free credits — enough to try Udio (or Suno) immediately. Then head to the Playground and switch the model selector from Suno to Udio.

For the full Udio reference, see the Udio model docs. For pricing context across third-party providers, see Suno API Pricing in 2026.

Last updated: 2026-06-04.