Real-song QA mindset
Engine changes are guarded by regression checks so improvements do not quietly damage known charts.
Accuracy
Chord detection is hard. The product should be honest about that.
Dense mixes, bass movement, borrowed chords, live timing, and fast passing changes can confuse every detector. GigaChords is built to improve with real evidence and keep the result editable.
Multiple analysis depths
Fast, Hearing, and Deep modes let the product balance speed and careful listening.
Editable, not arrogant
When the app gets a chord wrong, the workflow should help you fix the chart and keep moving.
What the engine listens for
A usable chord map is part audio evidence, part musical timing.
The GigaChords direction combines chroma evidence, chord templates, onset-aware smoothing, tempo grids, key-aware cleanup, and targeted repair passes. The public site should describe that at a human level, not drown musicians in diagnostics.
Website claim boundary
Say "clear starting chart" and "practice-ready timeline." Do not say "perfect transcription." That restraint makes the product more credible.
Get early access to the chord timeline app built for real practice.
Join the beta list and help shape the first public GigaChords builds.