Guide
User Guide
Getting Started
Click New practicefrom the library. You'll first choose which workflow fits the recording:
- Melody & solo notation - full transcription of a melody or solo line, with chords.
- Lick in every key - one short lick, automatically transposed through all 12 keys. See Lick in every key below.
Drop or browse for an MP3, WAV, or FLAC file, or a video (MP4, MOV) (max 100 MB). Processing runs through several stages - converting, separating instrument stems, transcribing notes, analyzing BPM/key, detecting chords, and building the notation - shown with a progress bar. Stem separation is the slowest step, typically 8-15 minutes on CPU. When it finishes you're returned to your library, where you can open the new practice item.
Account Access
If you need access to your account, automated password reset is not currently available. The immediate self-service option is to create a new account at /signup.
Once you've signed in again, your new practice items and settings will stay with the account you use.
The Practice View
Opening a song shows:
- A header with the song title, detected key, note count, and instrument picker
- The transport bar - playback, loop points, tempo, and (for songs) melody-source correction
- The notation staff, which auto-scrolls to follow playback
- Icons for downloading (⬇), the stem mixer (🎛, songs only), and notation settings (⚙)
Transport & Playback
⏮ ⏪ ▶/⏸ ⏩ - return to start, rewind/forward 5 seconds, play/pause.
♪ (Play notes)- plays the transcribed notes back as clean MIDI, separately from the actual audio. Useful for checking whether a transcription sounds right without the original recording's texture getting in the way.
A/B loop - click [A] and [B] at the points you want to loop between, then ↺ to turn looping on. ✕ clears the loop points.
Tempo slider - changes playback speed (0.25x-1.5x) without affecting pitch. This is different from the BPM ÷2 / ×2 buttons next to the detected tempo, which correct the song's detected BPM itself (for when the analyzer picks double-time or half-time by mistake) - that changes how the notation is laid out in measures, not how fast it plays back.
Instrument & Notation Settings
Click the instrument name in the header (e.g. "Tenor Sax ▾") to switch instruments - the notation re-transposes instantly without re-processing the audio. Currently supported: Tenor Sax, Alto Sax, Trumpet, Flute, Clarinet, Oboe.
The settings menu (⚙) toggles:
- Chord symbols above the staff
- Chord tones shown as gray noteheads alongside the melody
- Note stems & flags
- Sharp/flat accidental symbols
- Dark/light staff theme
Fixing the Melody Source
Songs are scored section by section to decide which stem (vocals, guitar, piano, or other) is carrying the melody at each moment - this is how the app follows a vocal verse into an instrumental solo and back. If it gets a section wrong, two tools can fix it:
- Lead from: Auto/Vocals/Guitar/Piano/Instrument - re-transcribes the entire song from one stem. Use only if the whole song should come from a single source; it discards the mixed per-section result.
- Lead source(expandable section) - lists each detected section and lets you pin a different stem for just that section, then apply. This is the one to use for "the solo at 1:10 should be guitar, the rest is right."
Editing Notation
Click a notehead to select it - a small red delete button appears above it. Click the button to remove that note. Click anywhere else on the staff to deselect without deleting. There is currently no undo, so a delete is immediate and permanent.
Exporting
Click the ⬇ icon in the header to download the current notation as MusicXML (for notation software like MuseScore or Finale) or MIDI. Both use written pitch - what's shown on screen and what you'd actually play on your instrument.
Lick in every key
Upload a short lick and it's laid out back-to-back in all 12 keys, cycling through the circle of fourths starting from the lick's own key. Each key gets a one-measure count-in (silent, with a metronome click in the audio) before the statement plays. Practice audio is pitch-shifted per key from your original recording, not re-synthesized.