A quality first-draft funscript from any audio track — in under a minute.
Phrase-level mode shaping
Each phrase is classified — tease, steady, fast, edging, break, chaotic — and the curve is sculpted accordingly. Edging builds 50→100, breaks go quiet, teases narrow. Not a uniform beat-thump.
Stable beat tracking
PLP-based tracker holds the beat through tempo changes and weak downbeats. Locked-BPM mode for tracks with a known fixed tempo. Sub-beat density: 1, 2, 4, or 8 actions per beat.
Comparison-strip preview
Every Style / Stroke-density / Tone option rendered as a velocity-coloured curve thumbnail with synced ◀▶ scroll and a 5-step zoom ladder (4s → 16s → 1m → 3m → 10m). Pick the option that fits before exporting.
FunscriptForge handoff
Save writes the .funscript plus an .analysis.json sidecar — chapter proposals, source metadata, generation parameters. FunscriptForge auto-loads it on funscript open. Generate fast, refine deliberately.
Stats sidebar
Mean position, range, average velocity, p90 velocity, peak velocity — the same metrics FunscriptForge surfaces. Compare a generated curve to a gold-standard reference at a glance.
Fully local
Runs on your machine. No cloud, no account, no telemetry. Your audio and funscripts never leave your computer. CLI and importable Python library for batch and pipeline workflows.
