Hear • Understand • Discover
Wavves
AI-powered music intelligence for song discovery, concert prep, and smart playlists.
Find a Song
Song candidates
Submit a lyric fragment or song description to understand the query and search Spotify for likely candidates.
AI Pipeline Explorer
How Wavves understood this song clue
Follow the same data as it moves from audio or text into model inference, structured intent, Spotify search, and ranked candidates.
Observed latency
Pending
STT model
Not run
Candidates
Pending
Capture
Input boundary
Wavves starts with either typed text or a browser audio recording that becomes the working song clue.
- Current clue
- It goes take me out tonight
- Audio size
- No recording yet
- Audio duration
- No recording yet
Speech-to-text
Hosted inference
Deepgram turns the audio waveform into text and returns model metadata, latency, confidence, and estimated cost.
- Provider
- Not run
- Model
- Not run
- Latency
- Pending
- Confidence
- Not reported
Intent
Query understanding
The intent layer converts the raw clue into structured fields that downstream services can use.
- Type
- Pending
- Lyrics
- Pending
- Model
- Pending
- Confidence
- Pending
Search
Downstream API
The Node app uses the structured intent to build a Spotify track-search query and records search latency.
- Spotify query
- Pending
- Latency
- Pending
- Total results
- Pending
Candidates
Ranking surface
Spotify candidates are shown with lightweight scores and explanations so the result is inspectable.
- Shown
- Pending
- Top result
- Pending
- Limitation
- Waiting for search
Find a Song
Can't remember a lyric? Hum it, sing it, or describe it.
Concert Prep
Generate a playlist from likely setlists.
Smart Playlists
Create playlists from natural language.
AI Pipeline Explorer
See exactly how the AI understood your request.
Workflow
Concert Prep
Status
Sign in, describe a concert request, and Wavves will build the playlist in the background.
AI Pipeline Explorer
Understand the path
Every AI feature should reveal how the request moved through preprocessing, model inference, downstream APIs, and the final music result.