Background music for video
Describe mood, tempo and instruments and get a track for the clip. Download the file and mix it with the video in your own editor.
CAPABILITY
AI music generator from a text description
The Sound section builds music from a text description: you set genre, mood and instruments, and get back an audio file. There are three modes — a track from a description, a song from your own lyrics, and an instrumental without vocals. The file format is chosen before the run.
- 3
- music generation modes
- 4
- file formats
- 3,500
- characters of lyrics
How it works
-
Pick a mode
A track from a description, a song from your own lyrics, or an instrumental without vocals. The mode decides which fields you fill in before the run.
-
Describe the track
Genre, mood, instruments and tempo go in as plain text, up to 2,000 characters. Separate fields let you state what should stay out of the track and set a seed for repeatability.
-
Choose a format and run it
MP3 is set by default; WAV, FLAC and OGG are available too. The cost is shown before the run; if your balance is too low, the request is not sent and the service shows how much is missing.
-
Collect the file
The finished track plays in the built-in player and downloads with one button. Every run stays in the section history: replay it, repeat the request or delete it.
What it does
- Input
- A text description of the track; in song mode, your own lyrics
- Modes
- Track from a description, song from your own lyrics, instrumental without vocals
- Track description
- Up to 2,000 characters
- Lyrics
- Up to 3,500 characters
- Negative prompt
- Up to 3,500 characters — what should stay out of the track
- Seed
- A whole number from 0 to 1,000,000
- File formats
- MP3 by default, WAV, FLAC, OGG
- Output per run
- One audio file
- Where to find it
- The Sound section in the web app and in the Telegram mini app, Sound mode in Studio
What people use it for
Instrumental under a voice-over
The no-vocals mode returns a track with no singing, for a podcast, a training video or a presentation.
A demo for lyrics you already have
Song mode takes your own lyrics: you can hear how the lines sit on a melody and compare several runs in different genres.
A reference for a brief
A rough track instead of a verbal description: show a client or a musician the direction rather than explaining it in words.
Models for this task
Full catalog
Sound
Fish Audio Drama 3
Fish Audio Drama 3: directed voice performances with plain-language directions in square brackets, two-voice dialogue
$0.0024
/ 1K chars
A minute of speech ≈ $0.002
Sound
HeyGen Voice Turbo
Faster, lower-cost HeyGen Voice with the same 10 voices
$0.009
/ 1K chars
A minute of speech ≈ $0.0076
Sound
HeyGen Voice
HeyGen text-to-speech: 10 ready voices for videos, presentations and avatars
$0.018
/ 1K chars
A minute of speech ≈ $0.015
Sound
ElevenLabs Turbo v2
Fast English-only ElevenLabs model
$0.024
/ 1K chars
A minute of speech ≈ $0.02
Sound
ElevenLabs Turbo v2.5
Low-latency ElevenLabs model balancing quality and speed, 32 languages
$0.024
/ 1K chars
A minute of speech ≈ $0.02
Sound
ElevenLabs Flash v2.5
Ultra-low-latency ElevenLabs model for conversational use, 32 languages
$0.024
/ 1K chars
A minute of speech ≈ $0.02
Sound
ElevenLabs Multilingual v2
Consistent ElevenLabs narration for long-form voice-overs and audiobooks, 29 languages
$0.048
/ 1K chars
A minute of speech ≈ $0.041
Sound
ElevenLabs Eleven v3
Emotionally rich ElevenLabs speech with inline audio tags, 70+ languages
$0.048
/ 1K chars
A minute of speech ≈ $0.041
Sound
ElevenLabs Eleven v4 Turbo
Faster Eleven v4: same expressive delivery and audio tags, 90+ languages
$0.024
/ 1K chars
A minute of speech ≈ $0.02
Sound
ElevenLabs Eleven v4
ElevenLabs' most expressive model: emotion and delivery via inline audio tags, 90+ languages
$0.048
/ 1K chars
A minute of speech ≈ $0.041
Sound
Fish Audio S2.1 Pro
Expressive speech you direct inside the text: emotions, pauses, two-voice dialogue; 80+ languages and voice cloning
$0.018
/ 1K chars
A minute of speech ≈ $0.015
Sound
Seed Audio 1.0
Expressive audio from a scene description: voice, ambience, background music and effects in one generation
$0.1896
/ min
What else it does
The things people usually discover only once they start working.
- A cost estimate is shown before the run; the final amount is fixed once the track is ready
- The charge comes off your balance for a successful run; a failed one is not billed
- Sound history lives in its own section, separate from the general file gallery
Frequently asked questions
What can I set about the track?
Genre, mood, instruments and tempo go in as plain text, up to 2,000 characters. Separate fields hold the negative prompt — what should stay out of the track, up to 3,500 characters — and the seed, a whole number from 0 to 1,000,000, for repeatability. The file format is chosen before the run, and the actual duration is shown on the finished track's card.
How do the three modes differ?
In the first mode the track is built from your text description. Song mode takes your own lyrics, up to 3,500 characters. Instrumental returns a track without vocals, for a podcast, a training video or a presentation. The mode decides which fields you fill in before the run.
What format does the file come in?
MP3 by default, with WAV, FLAC and OGG also available. The format is chosen before the run, and the downloaded file carries the matching extension.
What does one track cost?
The cost is shown before the run, but it is an estimate: the exact amount is fixed once the track is ready and can differ. It depends on the model you pick — current rates are on the pricing page.
What if the result is not what I wanted?
Run it again: repeat the request from the history or change the description. The seed handles repeatability, and the negative prompt takes out whatever got in the way.
Where are finished tracks kept?
In the Sound section history: replay a track in the built-in player, download it, repeat the request with the same settings or delete it. The section has its own history, separate from the general file gallery. The same section holds two more tools — text to speech and sound effect generation.
Related capabilities
Try it on your own task
A starting balance is already on your account, no card required. You pay only for what you generate.