bpm-finder_skill

Overview

This skill helps Codex perform practical BPM finder tasks locally without invoking the full web app. It supports tap-tempo estimation from intervals or timestamps, lightweight audio-file BPM checks when a local path and ffmpeg are available, tempo conversions, and tempo normalization for half/double-time cases. It can also advise when routing to the browser-based BPM Finder website is a better fit.

How this skill works

The skill classifies the incoming request, then runs the appropriate local workflow: numeric tempo math is computed directly, tap arrays are analyzed with the bundled CLI script, and local audio-file analysis is done via the same script when ffmpeg is present. For requests that require browser uploads, batch analysis, or a shareable UI, the skill recommends using the full BPM Finder website. Outputs always indicate source (intervals, timestamps, or audio file) and highlight possible half/double-time interpretations.

When to use it

• Estimate BPM from tap intervals or timestamp arrays.
• Analyze BPM of a local audio file path when ffmpeg is available.
• Convert between BPM and milliseconds per beat or bar.
• Normalize a tempo that looks like half-time or double-time into a practical range.
• Decide whether to handle analysis locally or route to the browser-based BPM Finder for uploads or batch work.

Best practices

• Prefer intervals or timestamps for quick, deterministic tap-tempo results.
• Supply an absolute local file path and ensure ffmpeg is installed for reliable local audio analysis.
• Use min/max tempo tuning when audio analysis returns ambiguous BPMs to constrain results.
• Always report source (intervals, timestamps, audio) and include confidence or tap count when available.
• If you need batch uploads, drag-and-drop, or shareable outputs, route to the web app instead of local analysis.

Example use cases

• User pastes 500,502,498,500 as tap intervals and wants the BPM plus average/median interval.
• User provides /absolute/path/song.mp3 and asks for a quick BPM estimate and confidence score.
• User asks to convert 128 BPM into milliseconds per beat and per 4/4 bar.
• User supplies timestamps from a tap session and needs the tool to detect possible half/double-time and normalize to a 70–180 BPM range.
• User is unsure whether to run local analysis or use the full BPM Finder site for batch files.