Frequently Asked Questions

Common questions about using MusicAPI.

Getting Started

How do I get started with MusicAPI?

Get your API key from the dashboard

Make your first API call

Subscribe to a plan whenever you're ready to go beyond the free quota

What's included in the free trial?

New accounts receive 75 free credits on signup — enough to evaluate the API before subscribing. For example:

3 Sonic generations (up to 6 songs, 20 credits per call)

Or 6 Producer generations (12 credits per call)

Plus utility calls like persona training, BPM, and stem separation at 1–30 credits each

No credit card required to start.

Music Generation

How do I create a song with my own lyrics?

Enable custom_mode and provide your lyrics in the prompt field:

{
  "custom_mode": true,
  "mv": "sonic-v4-5",
  "title": "My Song",
  "tags": "pop, rock",
  "prompt": "[Verse]\nYour lyrics here\n\n[Chorus]\nChorus lyrics here"
}

How do I let AI generate lyrics for me?

Set custom_mode to false and use gpt_description_prompt:

{
  "custom_mode": false,
  "mv": "sonic-v4-5",
  "gpt_description_prompt": "An upbeat song about summer adventures"
}

How do I create instrumental music without vocals?

Set make_instrumental to true:

{
  "custom_mode": false,
  "mv": "sonic-v4-5",
  "make_instrumental": true,
  "gpt_description_prompt": "A peaceful piano melody"
}

How do I control the singer's gender?

Use the vocal_gender parameter (only available for v4-5 and above):

{
  "custom_mode": true,
  "mv": "sonic-v4-5",
  "vocal_gender": "f",
  "prompt": "...",
  "tags": "pop"
}

Value	Description
`f`	Female vocal
`m`	Male vocal

Can I control the exact duration of a song?

Short answer: In Sonic, yes, in custom mode. Send duration (an integer, 10 to 360 seconds) in the request body — see Song Length (duration) in the Sonic Instructions. It is a target rather than a hard cut (a request for 45 typically lands at 43 to 45 seconds) and is accepted on create_music, persona_music, cover_music, cover_upload_music, extend_music and extend_upload_music. Credit cost does not change with the requested length.

It is only reliable in custom mode (you supply the lyrics). In our October 2026 tests on sonic-v6, 92–100% of custom-mode clips landed within 5 seconds of the target, but only 2–9% of AI description mode clips did: in description mode the model mostly picks its own length. If length matters, use custom mode, or use the tips below.

Nuro supports it too - see the nuro-instruction for details. Producer accepts length (1 to 240 seconds) and passes it to the model as a target - see the Producer instructions. We have not measured how closely Producer follows it, so treat it as a hint.

In AI description mode, you can use indirect methods to influence song length:

Model Version Limits

Different model versions have different maximum durations per generation:

Model Version	Max Duration
sonic-v3-5	~2 minutes
sonic-v4	~2 minutes
sonic-v4-5	~4 minutes
sonic-v4-5-plus	~4 minutes
sonic-v5	~4-8 minutes

Tips to Make Songs Shorter

Use ending tags - Add [Outro], [End], [Fade Out], or [Short Instrumental Outro] in your lyrics to signal the song should end:

[Verse]
Your verse lyrics...

[Chorus]
Your chorus lyrics...

[Outro]
Final words...
[End]

Write shorter lyrics - Less lyrics generally means shorter songs

Use concise structure - Fewer sections (e.g., just Verse → Chorus → Outro)

Tips to Make Songs Longer

Use the Extend feature - Generate a song, then use extend_music to add more content:

{
  "task_type": "extend_music",
  "continue_clip_id": "your-clip-id",
  "continue_at": 180,
  "custom_mode": true,
  "mv": "sonic-v4-5",
  "prompt": "[Bridge]\nNew section...\n\n[Final Chorus]\n..."
}

Write more lyrics - More content generally produces longer songs

Add more sections - Include multiple verses, pre-chorus, bridge, etc.

Use instrumental breaks - Add [Instrumental] or [Guitar Solo] tags

Extending & Covering Songs

How do I extend an existing song?

Use extend_music task type with the original clip_id and specify where to continue:

{
  "task_type": "extend_music",
  "continue_clip_id": "original-clip-id",
  "continue_at": 120,
  "custom_mode": true,
  "mv": "sonic-v4-5",
  "prompt": "[Bridge]\nNew lyrics for extension...",
  "tags": "pop"
}

How do I get the full extended song (original + extension)?

After extending, use concat_music to merge the original and extended parts:

{
  "task_type": "concat_music",
  "continue_clip_id": "extended-clip-id"
}

This produces a single full-length track combining both parts.

How do I create a cover version?

Use cover_music task type:

{
  "task_type": "cover_music",
  "continue_clip_id": "original-clip-id",
  "custom_mode": true,
  "mv": "sonic-v4-5",
  "prompt": "...",
  "tags": "jazz, smooth",
  "audio_weight": 0.6
}

The audio_weight parameter (0-1) controls how much the original influences the cover.

Uploading Your Own Music

How do I upload my own audio file?

Use the upload endpoint to get a clip_id:

Requirements:

Audio must be at least 6 seconds long (max 8 minutes)

Must be copyright-free or owned by you

Processing takes approximately 20 seconds

How do I extend/cover my uploaded music?

Use the special task types for uploaded music:

Operation	Task Type
Extend uploaded music	`extend_upload_music`
Cover uploaded music	`cover_upload_music`

Important: Do not use extend_music or cover_music for uploaded files. Use the _upload_ variants instead.

Personas (Virtual Singers)

What is a persona?

A persona is a virtual singer profile. Once created, you can use it to generate new songs in a consistent voice / style across many tracks.

Which endpoint should I use: `/persona` or `/create-voice`?

There are two ways to train a persona. Pick based on what you have as input:

Endpoint	Input	Best for	Durability
`POST /api/v1/sonic/persona`	An existing `clip_id` (from a prior Sonic generation or a `/sonic/upload` response)	Consistency across generations from a song you already have	Durable — trained on shared backend resources, not tied to a single private account
`POST /api/v1/sonic/create-voice`	A public HTTP(S) audio URL (MP3 / WAV) you host yourself	One-step voice cloning from your own recorded audio, when you don't want to upload through `/sonic/upload` first	Fragile — bound to a private backend account that may be retired, causing the persona to orphan without notice

Recommendation: default to /persona for production workloads. If your source is user-uploaded audio, the durable path is two steps — POST /api/v1/sonic/upload with the file, then POST /api/v1/sonic/persona with the returned clip_id. Reserve /create-voice for prototypes or scenarios where the one-step convenience outweighs the fragility risk.

How do I create a persona from a clip_id (durable)?

POST /api/v1/sonic/persona

{
  "name": "My Singer",
  "clip_id": "source-clip-id",
  "describe": "Smooth jazz vocal with warm tone",
  "styles": "jazz, smooth, warm"
}

The clip_id can come from any prior Sonic create_music response or from POST /api/v1/sonic/upload with your own audio file.

How do I create a persona from uploaded audio (one-step)?

POST /api/v1/sonic/create-voice

{
  "audio_url": "https://your-cdn.example.com/voice-sample.mp3"
}

For this endpoint, audio_url must be a public HTTP(S) URL whose pathname ends
with .mp3 or .wav (case-insensitive). Query strings and fragments are
allowed after the suffix.
This is a URL contract, not content inspection: the server does not download the
URL before submission, and the upstream service may still reject content that is
not actually MP3 or WAV.

Note: personas created this way are bound to private backend resources and may become unavailable without notice. When that happens the API returns {"type":"validation_error","error":"The persona is no longer available..."} and you'll need to re-train. For production workloads, prefer the /upload + /persona flow described above.

How do I use a persona to generate music?

Use persona_music task type with the persona_id:

{
  "task_type": "persona_music",
  "persona_id": "your-persona-id",
  "custom_mode": true,
  "mv": "sonic-v4-5",
  "prompt": "...",
  "tags": "pop"
}

Lyrics Generation

How do I generate lyrics?

Use the lyrics endpoint:

POST /api/v1/sonic/lyrics

{
  "prompt": "A love song about long distance relationships"
}

Cost: 2 credits per generation

Technical Questions

Does the API support Suno v6?

Yes. All three v6 models are selectable: sonic-v6, sonic-v6-wild and
sonic-v6-mini. See Sonic Instructions for what each one
is.

Suno released v6 on 2026-09-09 and retired v3.5 through v5.5 at the same time.
Older mv values are still accepted, but Suno runs them on v6, so the model you
name is no longer the model that runs. mv in the response still reports what
you asked for. A generation costs 20 credits on every model, including the
retired ids. See the Credits Usage Guide.

Model changes are announced on the API Changelog. Email
support@musicapi.ai with the subject changelog to get them by email.

What audio formats are supported?

Input: Varies by endpoint. /sonic/create-voice requires a public HTTP(S)
audio_url whose pathname ends in .mp3 or .wav; upload endpoints may
support additional formats documented by those endpoints.

Output: MP3 (audio_url), MP4 (video_url), WAV (via get-wav endpoint)

How long does generation take?

Operation	Typical Time
Create music	1-3 minutes
Extend music	1.5min
Cover music	1.5min
Concat music	1min
Upload music	~20 seconds

What's the recommended polling interval?

Poll every 15-25 seconds for task results.

Why did my request fail?

Common reasons:

Error	Cause	Solution
401 Unauthorized	Invalid API key	Check your API key
403 Forbidden	Insufficient credits	Purchase more credits
400 Validation Error	Invalid parameters	Check parameter values
429 Rate Limit	Too many requests	Wait 3 seconds between requests

Billing & Credits

How many credits does each operation cost?

See the Credits Usage Guide for complete pricing.

Quick reference:

Operation	Credits
Create/Cover	20
Extend	15
Concat	2
Upload	2
Persona/BPM/VOX	1

Do unused credits roll over?

Subscription credits: No, reset monthly

Extra credits: Yes, never expire

Which credits are used first?

Extra credits are consumed before subscription credits.

Output & Usage Rights

Are there watermarks on generated music?

No, all generated audio is watermark-free.

Can I use generated music commercially?

Yes, generated music can be used for commercial purposes. Check our terms of service for specific licensing details.

How many songs does each generation produce?

Most operations produce 2 songs per request (e.g. 20 credits for a create call), giving you options to choose from.

Still Have Questions?

Documentation: docs.musicapi.ai

Telegram: Join our channel for updates, examples, and support

Email: Contact us for enterprise inquiries: support@musicapi.ai

Music API FAQ & Troubleshooting

Frequently Asked Questions#

Getting Started#

How do I get started with MusicAPI?#

What's included in the free trial?#

Music Generation#

How do I create a song with my own lyrics?#

How do I let AI generate lyrics for me?#

How do I create instrumental music without vocals?#

How do I control the singer's gender?#

Can I control the exact duration of a song?#

Model Version Limits#

Tips to Make Songs Shorter#

Tips to Make Songs Longer#

Extending & Covering Songs#

How do I extend an existing song?#

How do I get the full extended song (original + extension)?#

How do I create a cover version?#

Uploading Your Own Music#

How do I upload my own audio file?#

How do I extend/cover my uploaded music?#

Personas (Virtual Singers)#

What is a persona?#

Which endpoint should I use: /persona or /create-voice?#

How do I create a persona from a clip_id (durable)?#

How do I create a persona from uploaded audio (one-step)?#

How do I use a persona to generate music?#

Lyrics Generation#

How do I generate lyrics?#

Technical Questions#

Does the API support Suno v6?#

What audio formats are supported?#

How long does generation take?#

What's the recommended polling interval?#

Why did my request fail?#

Billing & Credits#

How many credits does each operation cost?#

Do unused credits roll over?#

Which credits are used first?#

Output & Usage Rights#

Are there watermarks on generated music?#

Can I use generated music commercially?#

How many songs does each generation produce?#

Still Have Questions?#