PiAPI/DiffRhythm-Create Task

Overview

This endpoint from PiAPI's DiffRhythm API can generate audio based on prompt or reference audio with lyrics.
The lyric placement is highly flexible—PiAPI users can customize lyric integration using structured input formats. They can also choose different task-type depending on the target duration of the music they want.

Information&Pricing

Model	Task Type	Generate Audio Duration	Price (USD)
Qubico/diffrhythm	txt2audio-base	1.35 min	$0.02/generation
Qubico/diffrhythm	txt2audio-full	4.45 min	$0.02/generation

Example

To generate reasonable audio, Users could choose audio style(e.g., pop) based on text or upload Base64 or url of reference audio.
Here is an example that users could follow when they need to generate a song with lyrics:

[00:10.00] Drifting through the Milky Way’s glow, stars hum low and bright
[00:17.00] Every shadow hides a cosmos, burning without light
[00:24.00] You said you loved black holes’ silence, where time turns to dust
[00:31.00] I clung to supernova whispers, aching with their trust
[00:38.00] Orion’s belt held promises we strung like comet trails
[00:45.00] Now Saturn’s rings just echo how our gravity fails
[00:52.00] Your voice became a distant pulse, a quasars’ fading song
[00:59.00] I chart the void where constellations swore we’d belong
[01:06.00] Nebulas bleed hues we named—our scarlet, our chrome
[01:13.00] But cold dark matter cradles what the stars left alone…
[01:20.00] (Instrumental fade: synth waves & celestial chimes)

Users can control the musical style/tone through prompts when adding lyrics to precise time positions.

Request

Header Params

x-api-key

string

optional

your api key

Body Params application/json

model

enum<string>

required

Model name, currently select Qubico/diffrhythm.

Allowed value:

Qubico/diffrhythm

task_type

enum<string>

required

Currently could select txt2audio-base or txt2audio-full based on the length of audio that you want to generate.

Allowed values:

txt2audio-basetxt2audio-full

input

object

required

lyrics

string

optional

The lyrics of the audio following a specific structure (eg. [00:10.00] Drifting through the Milky Way’s glow, stars hum low and bright [00:17.00] Every shadow hides a cosmos, burning without light).

style_prompt

string

optional

Describe the style of audio that you want to generate.

style_audio

string

optional

Base64 or URL of the reference audio style.

config

object

required

webhook_config

object

optional

Example

{
  "model": "Qubico/diffrhythm",
  "task_type": "txt2audio-base",
  "input": {
    "lyrics": "",
    "style_prompt": "pop",
    "style_audio": ""
  },
  "config": {
    "webhook_config": {
      "endpoint": "",
      "secret": ""
    }
  }
}

Request samples

Shell

JavaScript

Java

Swift

PHP

Python

HTTP

Objective-C

Ruby

OCaml

Dart

Responses

🟢200Success

application/json

Body

model

string

optional

task_type

string

optional

input

object

optional

prompt

string

required

negative_prompt

string

optional

video

string

required

steps

integer

optional

seed

integer

optional

config

object

optional

webhook_config

object

optional

Example

{
  "model": "Qubico/diffrhythm",
  "task_type": "txt2audio-base",
  "input": {
    "lyrics": "",
    "style_prompt": "pop",
    "style_audio": ""
  },
  "config": {
    "webhook_config": {
      "endpoint": "",
      "secret": ""
    }
  }
}

Generate Audio

Overview#

Information&Pricing#

Example#

Request

Request samples

Responses

Overview

Information&Pricing

Example