OpenVoiceEverything. Release status

YOUR VOICE, TWO WAYS TO WORK

Less typing.
More doing.

Speak into the app you’re using. Or turn a whole folder of videos into text. OpenVoiceEverything brings both to your Mac.

Meet the CLI

Public release preparation in progress. Download not yet available.

macOS 15+ Apple silicon Personal Mac app

THE WORDS, WITHOUT THE TYPING
A little less busywork

DRAFT / PROJECT UPDATE

Hey team,

Let’s keep the next step simple.
Make room for the work
that matters.

Ready when you are.
Speak. Stop. It’s there.⌃⇧Space

Illustrated workflow · example shortcut, yours to choose

01Words where you need them 02Files your agent can work with 03A few steps to get started

01 / EVERYDAY DICTATION

Your shortcut.
Your train of thought.

Keep your cursor in the conversation, document, or draft. Use your chosen shortcut to dictate, then send the words straight to the field you were working in.

  1. 1

    Make the shortcut yours

    Choose a key combination or mouse button in Settings → Shortcuts → Primary dictation.

  2. 2

    Focus a field and speak

    Start recording with your shortcut. Choose how it behaves in Settings → Dictation.

  3. 3

    Stop, and carry on

    Clipboard Paste inserts the transcript and restores your previous clipboard after delivery. Keep a transcript copy as a backup if you like.

A RHYTHM THAT FITS YOU

Tap → Speak → Tap

Tap once to start, then tap again to finish. Keep both hands free while you speak.

Dictation needs Microphone access. Automatic paste also needs Accessibility permission. You’ll set these up in the app.

02 / A CLI YOUR AGENT CAN USE

A folder of videos.
A folder of possibilities.

Give your agent the folder. Give it a clear command. Get same-named transcripts it can search, summarize, and work with.

ove-agentLOCAL MAC · EXAMPLE
ove-agent transcribe ~/Videos -r --dry-run

See every input and its planned output before contacting the app or writing a file.

INPUT

Videos/

├─ interview.mp4

├─ notes.mov

└─ Field recordings/

└─ afternoon.wav

PLANNED OUTPUT

Videos/

├─ interview.txt

├─ notes.txt

└─ Field recordings/

└─ afternoon.txt

Illustrative file paths. Run these commands after the app and CLI are set up below.

Pick up where you left off

Existing outputs are skipped by default. Rerun a batch to retry failed files and leave completed transcripts in place.

Predictable for scripts

One JSON summary on stdout. Progress on stderr. Per-file results and clear exit codes tell your agent what happened.

Bring mixed media

Use native MP4/MOV and common audio formats. Optional ffmpeg adds MKV, WebM, AVI, and more; audio codec support varies.

TRY THIS WITH YOUR AGENT

“Transcribe everything
in this folder.”

Read the agent guide
Use the bundled OpenVoiceEverything CLI on this Mac to transcribe every audio/video in this folder recursively. Inspect help and settings, then show a dry-run plan. Save text beside each source with the same stem, preserve existing outputs, and report JSON counts plus failed inputs.

03 / START WITH YOUR MAC

A little setup.
A lot less typing.

The planned installer is for Apple silicon Macs running macOS 15.0 or later. You use the CLI on the same Mac as the running app.

01 Set up the app

Open OpenVoiceEverything, finish onboarding, and select a speech model. Some models need a separate download. Grant Microphone and Accessibility permissions for dictation and automatic paste.

For file batches, your agent needs readable media paths on the same Mac. The CLI does not need to record your microphone.

02 Enable the bundled CLI

The CLI needs Python 3.9 or later. Run the bundled helper to enable the app’s local API, then check it is reachable:

"/Applications/OpenVoiceEverything.app/Contents/Helpers/ove-agent" ensure
"/Applications/OpenVoiceEverything.app/Contents/Helpers/ove-agent" health

If the app was already running with its API disabled, use ensure --restart. Model preparation is separate from API health.

To use the shorter commands shown above:

mkdir -p "$HOME/.local/bin"
ln -s "/Applications/OpenVoiceEverything.app/Contents/Helpers/ove-agent" "$HOME/.local/bin/ove-agent"
export PATH="$HOME/.local/bin:$PATH"

Add the PATH line to your shell configuration if needed. The symlink command leaves an existing command untouched.

03 Preview, then transcribe

Inspect ove-agent settings and confirm the selected model is installed. Use --dry-run to review paths, then run the batch. Existing transcripts are skipped.

Same-stem files such as clip.mp4 and clip.wav become clip.mp4.txt and clip.wav.txt when they would share an output. Ambiguous destinations across input roots fail before transcription.

Optional ffmpeg on your PATH expands format support. Use --ffmpeg always if a native container has an unsupported audio codec.

OpenVoiceEverything app icon

IN PREPARATION

Make room for
your next thought.

Public release preparation in progress. Download not yet available.

Personal installer · macOS 15+ · Apple silicon

The personal installer is ad-hoc signed, not notarized. Expect a macOS security warning and manual updates.

Licensed under GNU GPLv3. A corresponding source archive will accompany the installer.