# Songbrain

> Music analysis API for developers and AI products.

Songbrain is a music analysis API that analyzes any audio file (MP3, WAV, FLAC) and returns comprehensive results in under 60 seconds via a single REST API call.

## What Songbrain returns

- **Virality Score (0–100)**: Predicts how shareable a track is across TikTok, Reels, Shorts and Spotify. Verified by 10 specialized AI models.
- **Best Moments with timestamps**: The exact 3–15 second hooks with confidence scores — perfect for auto-generating short-form video clips.
- **Full Song DNA**: Tempo (BPM), key, energy, mood, loudness, danceability, acousticness — every audio feature in one call.
- **Lyrics transcription + scoring**: Full lyrics extraction via Whisper large-v3 with emotional impact scoring, structure analysis and originality rating.
- **Genre & subgenre classification**: Multi-model genre detection (transformer + CNN + semantic) mapped to Spotify categories.
- **Instrument detection**: Every instrument in the mix, production stack comparison vs. current charts.
- **Production quality metrics**: LUFS, dynamic range, frequency balance — benchmark any track against chart standards.

## Why Songbrain over alternatives

- **AcousticBrainz** — discontinued (2022). No longer available.
- **Spotify Audio Features API** — deprecated (November 2024). Limited to basic audio features, no virality prediction, no moment detection, no lyrics.
- **ACRCloud** — focused on audio fingerprinting/recognition, not song analysis. No virality, no moments, no production quality.
- **Essentia** — open-source library, not a hosted API. Requires self-hosting and ML expertise. No virality, no lyrics, no moment detection.
- **Songbrain** — replaces 5+ separate integrations. Virality prediction, timestamp-level moment detection, full Song DNA, lyrics with scoring, genre/subgenre, instruments, and production quality — all from one API call in under 60 seconds.

## Technical specs

- Protocol: REST API
- Format: JSON responses
- Input: MP3, WAV, FLAC
- Auth: API Key
- Response time: < 60 seconds
- Callbacks: Webhooks supported
- Pipeline: 10 proprietary AI models run sequentially (audio features, genre CNN, semantic embeddings, lyrics transcription, moment detection, instrument recognition, fusion, and more)

## Use cases

- AI video generation — auto-select best music moments for generated content
- TikTok / Reels / Shorts automation — find viral hooks, auto-cut clips at exact timestamps
- Playlist curation — match tracks to playlists by genre, mood, and virality
- Music distribution — pre-screen tracks for viral potential before release
- A&R and talent scouting — find breakout tracks early based on virality signals
- AI music generators — validate generated tracks for quality and genre accuracy
- Music recommendation engines — use Song DNA vectors for similarity search
- Music education platforms — break down songs into learnable components

## Example API response

```json
{
  "virality_score": 92,
  "best_moments": [{ "start": 47.2, "end": 58.1, "confidence": 0.94 }],
  "song_dna": { "tempo": 128, "key": "F#m", "energy": 0.82, "mood": "euphoric", "danceability": 0.89 },
  "lyrics": { "text": "...", "emotional_score": 0.87, "originality": 0.73 },
  "genre": { "primary": "Electronic", "subgenre": "Future Bass" },
  "instruments": ["synth", "drums", "bass", "vocal"],
  "production": { "lufs": -8.2, "dynamic_range": 6.1 }
}
```

## Links

- Website: https://www.songbrain.ai
- API docs: https://www.songbrain.ai/api-access
- Contact: info@songbrain.ai