Riffusion API

Code for API access to Riffusion's frontier models to make incredible music

This API is in a private beta and subject to change. Contact api@riffusion.com for inquiries and access.

Usage

Set your API key:

export RIFFUSION_API_KEY="<your-api-key>"

Install the Python client:

pip install riff-api

Make some music:

from riff_api import RiffAPIClient

client = RiffAPIClient()

response = client.prompt(
    "Explain the concept of time in French, piano chill",
    save_to="chill.m4a",
)

This will generate a full song with lyrics and music. Run multiple times for variations.

Endpoints

/prompt

This endpoint creates a song from a single natural language description the desired lyrical content and/or musical style. It's the simplest way to create.

The API will generate lyrics based on your topic, as well as pick specific sound prompts. If you don't describe a musical style or lyrics content, it will choose for you.

Get creative with your topics! Here are a few ideas:

• "Rap fun facts about Alaskan history"
• "Explain the concept of time in French"
• "My nephew Remi is a superhero with laser eyes. Make him a theme song with a rock orchestra"

Currently the API only supports returning base64 encoded bytes of an m4a file.

Both /prompt and /compose accept an optional model argument to select the model to use. The default is "FUZZ 2.0".

/compose

This endpoint provides a more powerful capability for music lovers to craft the exact sound they want. You can specify custom lyrics and multiple sound prompts with individually controllable strengths for deeper control.

The returned output contains detailed timestamps for each word in the lyrics.

/models

This endpoint returns the list of available models based on your API key.

Repo

• riff_api/types.py - API schema
• riff_api/examples/prompt - Examples for creating a song via the client, requests, and curl
• riff_api/examples/compose - Examples for custom creation