songsee

steipete/clawdis

Generate spectrograms and multi-panel audio visualizations from MP3, WAV, and other formats.

What is songsee?

songsee is a CLI tool that creates visual representations of audio files, including spectrograms and feature panels (mel, chroma, HPSS, self-similarity, loudness, tempogram, MFCC, flux). Use it to analyze audio characteristics, debug signal processing, or create publication-ready visualizations.

Generate spectrograms from audio files
Create multi-panel visualizations combining multiple audio features
Extract and visualize time slices from longer tracks
Support multiple visualization types: spectrogram, mel, chroma, HPSS, self-similarity, loudness, tempogram, MFCC, flux
Customize output with different color palettes (classic, magma, inferno, viridis, gray)
Read audio from stdin and output to JPG or PNG format

How to install songsee

npx skills add https://github.com/steipete/clawdis --skill songsee

Prerequisites

songsee CLI tool (installable via brew: steipete/tap/songsee)
ffmpeg (optional, for non-WAV/MP3 formats)

Claude Code

Cursor

Windsurf

Cline

How to use songsee

1.Install songsee using brew: brew install steipete/tap/songsee
2.Run basic spectrogram: songsee track.mp3
3.For multi-panel visualization, add --viz flag with comma-separated features: songsee track.mp3 --viz spectrogram,mel,chroma,hpss
4.Customize output size with --width and --height flags
5.Specify time range with --start and --duration for slicing
6.Choose output format (jpg or png) with --format and filename with -o

Use cases

Good for

Analyze frequency content and spectral characteristics of music or speech
Create feature-rich visualizations for music information retrieval research
Debug audio processing pipelines by inspecting multiple feature representations simultaneously
Generate publication-quality figures for academic papers on audio analysis
Extract and visualize specific sections of long audio files for detailed inspection

Who it's for

Audio engineers and signal processing researchers
Music information retrieval (MIR) practitioners
Machine learning engineers working with audio data
Acousticians and sound designers
Academic researchers publishing audio analysis work

songsee FAQ

What audio formats does songsee support?

WAV and MP3 are decoded natively. Other formats require ffmpeg to be installed.

Can I visualize just a portion of an audio file?

Yes, use --start (in seconds) and --duration (in seconds) flags to extract and visualize a specific time slice.

What visualization types are available?

spectrogram, mel, chroma, hpss, selfsim (self-similarity), loudness, tempogram, mfcc, and flux. Combine multiple with --viz as comma-separated values.

How do I change the color scheme?

Use the --style flag with options: classic, magma, inferno, viridis, or gray.

Can I read audio from stdin?

Yes, use - as the filename (e.g., cat track.mp3 | songsee - --format png -o out.png).

Full instructions (SKILL.md)

Source of truth, from steipete/clawdis.

name: songsee description: "Generate spectrograms and feature-panel visualizations from audio with the songsee CLI." homepage: https://github.com/steipete/songsee metadata: { "openclaw": { "emoji": "🌊", "requires": { "bins": ["songsee"] }, "install": [ { "id": "brew", "kind": "brew", "formula": "steipete/tap/songsee", "bins": ["songsee"], "label": "Install songsee (brew)", }, ], }, }

songsee

Generate spectrograms + feature panels from audio.

Quick start

Spectrogram: songsee track.mp3
Multi-panel: songsee track.mp3 --viz spectrogram,mel,chroma,hpss,selfsim,loudness,tempogram,mfcc,flux
Time slice: songsee track.mp3 --start 12.5 --duration 8 -o slice.jpg
Stdin: cat track.mp3 | songsee - --format png -o out.png

Common flags