Forgot The Words - API Backend

Backend API for "I Forgot The Words To This Song" - removes vocals from songs so you can sing your own version.

Powered by Meta SAM Audio.

API Endpoints

/separate_audio

Separates audio based on text description.

Parameters:

• audio_path: Audio file
• description: What to isolate (e.g., "singing voice, vocals")
• predict_spans: Auto-detect timing (default: true)
• reranking_candidates: Quality setting (default: 1)

Returns: [target_audio, residual_audio]

Usage

from gradio_client import Client

client = Client("neonwatty/forgot-the-words-api")
result = client.predict(
    audio_path="song.mp3",
    description="singing voice, vocals, human voice",
    predict_spans=True,
    reranking_candidates=1,
    api_name="/separate_audio"
)

vocals, instrumentals = result