Seedance 2.5 from ByteDance is officially available on Apiframe.

Mureka API Guide: AI Music Generation, Pricing & Code (2026)

Mureka API explained: access, features, real pricing, and working code for AI music in 2026.

Janice Last updated November 12, 2026 August 10, 2026 · 7 min read Beginner
Mureka API Guide: AI Music Generation, Pricing & Code (2026)

Mureka has been quietly building a reputation as the fast, budget-friendly option in AI music generation, the one you reach for when you need a lot of tracks without much setup. It does not get the same attention as Suno or Udio, but it has a real developer API behind it with published per-song rates. Here is what it does, what it costs, and how to call it.

What Is Mureka

Mureka is an AI music generation platform that turns text prompts or full lyrics into finished songs, complete with vocals and instrumentation, or instrumental-only tracks when you do not want vocals. It sits in the same category covered in our broader AI music generation API guide, and if you are still deciding whether to build on a music API at all, our complete AI music API guide covers the fundamentals first.

Mureka gets compared directly against Suno and Udio in our AI music API comparison. The short version: Mureka trades some vocal polish for speed and a lower cost per track, which makes it a strong pick for high-volume or budget-sensitive projects rather than a flagship consumer music product. For a closer look at the alternatives, see our Suno API guide and Udio API guide.

Songs can run up to 5 minutes 30 seconds, and a single request returns up to 3 songs (2 by default). Instrumental background tracks are capped slightly shorter, at 4 minutes 30 seconds. Mureka currently generates lyrics in Chinese, English, Japanese, Korean, Portuguese, Spanish, German, French, Italian, and Russian.

Mureka API Access and Authentication

Mureka's developer API is a separate product from its consumer web app. Signing up on the Mureka website does not give you API access, and the reverse is also true. The two are billed and managed independently, so your website membership and credits will not carry over.

To get API access:

  1. Sign up for an account on the Mureka developer platform at platform.mureka.ai.
  2. Open the "API Keys" tab in your dashboard to create and manage keys.
  3. Fund your account. Mureka's API runs on a prepaid top-up balance rather than a subscription (more on that below), and the entry-level $10 trial can only be bought once per account.

Every request goes to https://api.mureka.ai and needs your key in the Authorization header as a bearer token:

bash
curl https://api.mureka.ai/v1/song/generate \
  -H "Authorization: Bearer $MUREKA_API_KEY" \
  -H "Content-Type: application/json"

Keep the key on your server. Mureka's own docs are explicit that it should never be exposed in browser or client-side code, since anyone who can read your page source could then spend your balance. Our guide to AI media API security best practices covers how to proxy requests safely.

Mureka API Features

Mureka's API is split into several endpoints rather than one catch-all "generate" call, which is worth knowing before you start integrating:

  • Lyrics to song (POST /v1/song/generate): generate a full song when you supply your own lyrics.
  • Prompt to song (POST /v1/song/easy-generate): generate a full song from a style prompt alone, letting Mureka write the lyrics.
  • Generate instrumental (POST /v1/instrumental/generate): instrumental-only tracks, no vocals. Mureka's pricing page lists this as BGM, short for background music.
  • Generate lyrics / Extend lyrics: standalone lyric generation and extension, useful if you want to review or edit lyrics before generating audio. Priced separately at $0.009 per set of lyrics or $0.002 per single line.
  • Soundtrack: upload an image or video clip and Mureka scores it, matching the mood and pacing of what it sees.
  • Extend song, Remix, Region editing, Stem export, Vocal cloning, Music transcription, and a few other endpoints for working on tracks you have already generated. Region editing is the handy one: it rewrites a 10 to 30 second window of a song without regenerating the whole thing.

For genre, mood, and vocal style, you steer generation with a style prompt (for example, "r&b, slow, passionate, male vocal"), the same approach Suno and Udio use.

Each request also takes a model parameter. You can pass auto and let Mureka pick, but the model you choose directly changes what you pay, so it is worth setting explicitly once you know which one suits your output. Mureka also has its own text-to-speech endpoint at $4.90 per hour of audio, though if voice is central to your product our ElevenLabs API guide covers the dedicated options.

Mureka API Pricing

Mureka publishes its API rates openly, priced per song rather than per request. Prices vary by endpoint and model:

  • Lyrics to song: $0.03 per song on V7.6, $0.045 on O2, V8, and V9, and $0.15 on V9.5.
  • Prompt to song: $0.30 per song on V7.6, O2, V8, and V9, and $0.50 on V9.5.
  • Instrumental (BGM): $0.03 per track on V7.6, $0.045 on V8 and V9, and $0.15 on V9.5.
  • Soundtrack (image or video to music): $0.10 per song, the same rate across V7.6, V8, V9, and V9.5.
  • Lyrics only: $0.009 per set of lyrics, or $0.002 per line.
  • Lyrics video: $0.10 per video.
  • Vocal cloning: $5 per voice. That is far above everything else on this list, but it is a one-time cost: the resulting Vocal ID never expires and can be reused across unlimited generations.
  • Stem export: $0.06 per song for five stems (vocals, drums, bass, accompaniment, and everything else), $0.20 for a vocal and instrumental split, or $0.70 for up to 12 separate instrument tracks.

The gap between the two main endpoints is the thing to plan around. Prompt to song costs ten times what lyrics to song costs on the same model, because Mureka is writing the lyrics for you. The arithmetic is worth doing: generating lyrics separately at $0.009 and then feeding them to the lyrics-to-song endpoint at $0.03 comes to $0.039 a track, against $0.30 for the same thing in one call. At volume that is the difference between a $39 bill and a $300 one for 1,000 tracks.

A few billing rules shape how you budget:

  • Access runs on a prepaid top-up balance, not a recurring subscription. You buy credit and it is consumed as you generate.
  • Your balance stays valid for 12 months from your last recharge, and any new recharge extends the validity of your existing balance too.
  • Credits are consumed oldest first, and paid credit is always spent before bonus credit.
  • How many tracks you can generate at once depends on the size of a single top-up: $10 gets you 1 request at a time, $1,000 gets 5, $3,000 gets 15, $5,000 gets 25, and $30,000 gets 150.
  • Larger top-ups currently come with bonus credit (5% at $3,000, 10% at $5,000, 20% at $30,000), though Mureka lists this as a limited promotion rather than standing policy.
  • Downloading generated songs is free. Exporting the separated stems costs extra, at the rates above.
  • All credits are non-refundable, both purchased and bonus, once spent or once the validity window lapses.

One rule catches people out: concurrency is tied to each individual purchase, not your running total. Buying $1,000 of credit five separate times still leaves you at 5 simultaneous requests, not 25. If throughput matters, you need one larger top-up rather than several small ones, which is a meaningful difference in how you plan a budget.

Those concurrency tiers are the real constraint for most teams. At the entry tier you are limited to one generation at a time no matter how much credit you have left, so throughput, not price, is usually what pushes projects to a higher tier. If speed and throughput are central to your build, our unified media API latency guide goes deeper on how to plan around that.

For a wider view of how these rates stack up against other providers, see our breakdown of AI music API pricing in 2026. If you would rather have one flat rate across models and endpoints, Apiframe prices Mureka generations at 7 credits ($0.07) per song, which lands above Mureka's cheapest lyrics-to-song rate but well below its prompt-to-song rate. Our guide on how to choose an AI media API walks through that kind of trade-off in more detail.

Code Example: Generating a Track with Mureka

Mureka does not generate tracks instantly. You send a request, get back a task ID and a preparing status, then keep checking a query endpoint until the track is ready. Here is how that looks in Python, and our step-by-step music API tutorial walks through the same pattern end to end.

python
import requests
import time

API_KEY = "YOUR_MUREKA_API_KEY"
headers = {
    "Authorization": f"Bearer {API_KEY}",
    "Content-Type": "application/json",
}

# 1. Submit the generation request
submit = requests.post(
    "https://api.mureka.ai/v1/song/generate",
    headers=headers,
    json={
        "lyrics": "[Verse]\nIn the stormy night, I wander alone\nLost in the rain, feeling like I have been thrown",
        "model": "auto",
        "prompt": "r&b, slow, passionate, male vocal",
    },
)
task = submit.json()
task_id = task["id"]

# 2. Check back until the task finishes, one way or the other
query_url = f"https://api.mureka.ai/v1/song/query/{task_id}"
while True:
    task = requests.get(query_url, headers=headers).json()
    if task["status"] in ("succeeded", "failed"):
        break
    time.sleep(5)

print(task)

Watch which status you wait on. A task moves through preparing, queued, and running before it reaches succeeded or failed. Break on one of those two finished states, not simply when preparing clears, or your loop will exit while the track is still being made. Checking on a timer like this works fine at low volume, but at scale you want Mureka to call you instead. Our guide on setting up webhook callbacks for media generation covers that pattern.

If you would rather skip writing lyrics yourself, swap the request to POST /v1/song/easy-generate and drop the lyrics field, keeping the price difference in mind. For instrumental-only output, use POST /v1/instrumental/generate instead.

Error responses return an HTTP status of 400 or higher with a body like {"error": {"message": "..."}, "trace_id": "..."}. The ones to handle:

  • 400 for malformed request parameters.
  • 401 for a bad or missing key.
  • 403 if you are calling from an unsupported region.
  • 429 for two different problems: sending requests faster than your concurrency limit allows, or running out of credit. The message text tells you which.
  • 500 for a server-side error on Mureka's end.
  • 503 when the generation engine is overloaded. Retry after a short wait.

The double meaning of 429 is worth handling carefully. Retrying a rate limit makes sense, but retrying an empty balance just burns requests, so read the error message before deciding whether to try again rather than reacting to the status code alone.

Mureka vs Suno vs Udio: Quick Take

Mureka's core trade-off is speed and cost against vocal polish. It generates tracks faster and, at $0.03 to $0.045 per song for lyrics-to-song, runs well below what Suno and Udio typically charge, but vocal quality is noticeably behind the top two in refinement.

That makes it a reasonable default for high-volume, budget-sensitive work like background music for short-form video at scale, or soundtracks for indie games, and a less obvious choice for a consumer-facing product where vocal quality is the main selling point.

For the full breakdown on licensing, genre range, and audio quality across all three, see our AI music generation API comparison. If you would rather reach all three models through one integration instead of juggling separate accounts, start with what a unified AI media API actually is, then the unified AI API guide, or browse our roundup of the best AI media generation APIs.

FAQ

Is there an official Mureka API?

Yes. It is a separate product from the consumer Mureka web app, run through platform.mureka.ai with its own account, API keys, and prepaid billing. Your website membership and credits do not carry over to it.

How much does Mureka cost per track?

Lyrics to song starts at $0.03 per song on V7.6 and runs to $0.15 on V9.5. Prompt to song is significantly more at $0.30 to $0.50 per song, since Mureka writes the lyrics. Instrumentals start at $0.03 per track. Access runs on a prepaid balance you top up, and a larger single top-up unlocks more simultaneous generations.

Can Mureka generate instrumentals only?

Yes, through the dedicated POST /v1/instrumental/generate endpoint, separate from the vocal song generation endpoints. Pricing starts at $0.03 per track, and instrumental tracks can run up to 4 minutes 30 seconds.

Is Mureka good for commercial use?

According to Mureka's own FAQ, content generated through paid API calls comes with full commercial usage rights, covering products, platform distribution, advertisements, and video. Check Mureka's terms of service for specifics before shipping, and see our guide on copyright and commercial rights for AI-generated content for the wider legal picture.

How long does generation take?

Mureka does not publish a fixed turnaround time, and the API hands back a task rather than a finished track. You submit the job, then check GET /v1/song/query/{task_id} until the status reaches succeeded or failed.

What languages does Mureka support?

Chinese, English, Japanese, Korean, Portuguese, Spanish, German, French, Italian, and Russian for lyrics. Note that the Extend Song endpoint supports a narrower set (English, Spanish, Portuguese, Chinese, Japanese, and Korean), and other languages may throw errors there.

You can also get an API key and start with free credits if you want Mureka alongside Suno, Udio, and dozens of other models with one predictable, published credit cost. If you want to compare free allowances first, see our comparison of AI media APIs with free trials.

Power your next AI product with Apiframe.

Instant access to 70+ media models through a single API. Start free and scale when you're ready.