Drop audio onto WAV, MP3, AAC / M4A or FLAC. Originals stay intact; existing results are never overwritten.
Details and limitations
WAV: 16 / 24 bit, 44.1 / 48 / 96 kHz. MP3: 192 / 256 / 320 kbps. AAC / M4A: 256 kbps. FLAC compresses losslessly, but cannot restore detail lost in an MP3.
NOTES AND PARTS
From audio to editable notes.
Basic Pitch Audio → MIDI transcribes melodies and chords. Edit notes and velocity in Piano Roll, with Undo / Redo and export at hand.
Details and limitations
Piano Roll: pitch, length, velocity and group selection. MIDI Split preserves timing and events in separate parts. Render MIDI to audio. Ghost Notes helps review hypotheses. No drum transcription or automatic instrument detection; Electric Piano is the default sound.
LOCAL STEMS · BETA
Listen to your recording in parts.
STEMS creates vocals, drums, bass and other sounds as separate audio files. VOCALS and DRUMS extract a selected audio stem.
Details and limitations
Separation is approximate: bleed and artifacts are possible. UMX is the default after installation; Kim and SCNet are experimental. Guitar / Keys are unsupported. DRUMS does not create MIDI.
MEDIA POOL AND QUEUE
A result becomes your next step.
Follow progress and cancel tasks. Saved results appear in Media Pool: audition a file, drag it to another tile, or reveal it in Finder.
Details and limitations
Removing a Media Pool reference keeps the file. Output folder and settings are shared across conversion workflows. The Media Pool capture uses test files.
FINDER · OPTIONAL
Another format. Keep your context.
Select audio in Finder → Control-click → Services → an AudioFlow command. One file or several enter the shared queue, with progress and cancellation in the menu bar.
Details and limitations
Explicitly install the add-on: Settings → Finder Quick Conversion → Install / Update. No models are needed. Originals stay intact. Full setup instructions and exact command names follow below.
Select one or more audio files and convert them through the system Services menu. You do not need to open the main AudioFlow window.
01
Select your files
One track or several audio files in Finder.
02
Open Services
Control-click or right-click → Services.
03
Choose your format
An AudioFlow command adds the files to the shared conversion queue.
WAV · MP3 · AAC / M4A · FLAC
Keep your creative flow
Progress and Cancel are available in the menu bar. Results use your shared AudioFlow settings. Originals are preserved and existing files are never overwritten.
This diagram shows the steps, not the macOS interface.
Enable it once in Settings
Settings → Finder Quick Conversion (optional) → Install / Update. Installation links the add-on to this copy of AudioFlow. No models or downloads are required.
If commands are hidden: System Settings → Keyboard → Keyboard Shortcuts → Services. macOS controls their placement and enabled state.
Exact command names in macOS
AudioFlow — Convert to WAV
AudioFlow — Convert to MP3
AudioFlow — Convert to AAC (M4A)
AudioFlow — Convert to FLAC
These commands live under Services, not in a separate top-level Finder menu.
YOUR FILES STAY WITH YOU
Local. And on your terms.
01
No audio uploads
Conversion and processing run on your Mac. AudioFlow does not send your audio or MIDI to the cloud.
02
Install models once
Install optional components explicitly in Settings. Compatible app updates reuse the models you already have.
Yes. Project support is voluntary and is not required to access features. There is no public download for the closed beta yet.
Do I need internet for processing?
Audio and MIDI are processed locally. Internet is needed for the first download of optional components and explicitly requested updates. These downloads contact model distribution servers; they do not upload your audio.
Do models download again after each app update?
No. Optional models are stored in your per-user components folder, separately from the app. Compatible versions reuse that installation. Install, update and repair actions are explicit in Settings.
Can I turn drums into MIDI?
No. Audio → MIDI is for pitched audio: melodies and chords. DRUMS separates a drum audio stem. Automatic instrument, tempo and key detection are not available.
How clean are the stems?
This is a beta feature and quality depends on the recording. Instrument bleed and artifacts are possible: keep your originals and listen to the results. UMX is the default after installation. Experimental Kim and SCNet are separately selected; Kim is limited to 32 seconds. Studio-quality results are not guaranteed.
What if macOS blocks the app?
This beta is ad-hoc signed and not notarized. If you trust the sender and have verified the supplied checksum, “Open Anyway” in Privacy & Security may be available after a launch attempt. Follow Apple’s guidance linked above. Do not disable Gatekeeper. This does not guarantee launch on every Mac.
How can I give useful beta feedback?
Include the version and build, your macOS version, the action you took and the observed result. Only share audio or private files if you choose to. Do not send passwords or other secrets.
Your music. Your choice to support.
If AudioFlow helps your work, you can support its development.