POST
/
api
/
v1
/
client
/
txt2music
Text to Music
curl --request POST \
  --url https://api.modelbeam.srv1069417.hstgr.cloud/api/v1/client/txt2music \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: multipart/form-data' \
  --form 'caption=<string>' \
  --form model=AceStep_1_5_Turbo \
  --form 'lyrics=<string>' \
  --form duration=305 \
  --form inference_steps=50 \
  --form guidance_scale=10 \
  --form seed=123 \
  --form format=flac \
  --form bpm=165 \
  --form 'keyscale=<string>' \
  --form timesignature=2 \
  --form 'vocal_language=<string>' \
  --form reference_audio='@example-file' \
  --form 'webhook_url=<string>'
{
  "data": {
    "request_id": "550e8400-e29b-41d4-a716-446655440000"
  }
}

Authorizations

Authorization
string
header
required

API key obtained from the ModelBeam dashboard

Body

multipart/form-data
caption
string
required

Music description / style prompt

Required string length: 1 - 2000
model
string
required
Example:

"AceStep_1_5_Turbo"

lyrics
string
required

Song lyrics. Use "[Instrumental]" for no vocals

duration
number
required

Duration in seconds

Required range: 10 <= x <= 600
inference_steps
integer
required

Inference steps (8 for turbo, 32+ for base)

Required range: 1 <= x <= 100
guidance_scale
number
required

Guidance scale

Required range: 0 <= x <= 20
seed
integer
required

Random seed (-1 for random)

format
string
required

Output format

Example:

"flac"

bpm
integer

Beats per minute

Required range: 30 <= x <= 300
keyscale
string

Musical key and scale, e.g. "C major"

timesignature
enum<integer>

Time signature

Available options:
2,
3,
4,
6
vocal_language
string

Vocal language code

reference_audio
file

Reference audio for style transfer (max 10MB)

webhook_url
string
Maximum string length: 2048

Response

Music generation job created

data
object