Skip to content

SHARE.md

Latest commit

 

History

History
85 lines (66 loc) · 3.42 KB

SHARE.md

File metadata and controls

85 lines (66 loc) · 3.42 KB
Raw

Sharing Echo with a non-technical friend

Echo is free, local, and unsigned (no Apple Developer account). With the steps below your friend gets a working app without installing any developer tools — the only manual step is a one-time Gatekeeper bypass (unavoidable without paid Apple notarization).

1. Build the shareable app (on YOUR Mac)

You need, one time: Rust (curl https://sh.rustup.rs -sSf | sh), Xcode Command Line Tools (xcode-select --install), cmake + git (brew install cmake git), and Node.

npm install
bash scripts/package-mac.sh

This compiles the native helpers, builds whisper-cli, bundles them into the app, and produces a small (~25 MB) DMG:

src-tauri/target/release/bundle/dmg/Echo_0.1.0_<arch>.dmg

The helpers + whisper-cli are baked in — on first launch the app copies them into ~/Library/Application Support/echo, so your friend needs no Xcode tools, Homebrew, or git/cmake.

The Whisper model (~142 MB) is not bundled — it keeps the DMG small. On first launch, onboarding downloads it once (Setup Whisper → the model lands in ~/Library/Application Support/echo/models). So your friend does need internet the first time they set up Whisper. After that, transcription is fully offline.

Chip match: this .dmg is built for your Mac's chip (Apple Silicon or Intel) and runs on a friend with the same chip. For a universal build that runs on both, push a vX.Y.Z git tag and use the GitHub release workflow instead (.github/workflows/release.yml).

Optional sanity check before sending:

plutil -p "src-tauri/target/release/bundle/macos/Echo.app/Contents/Info.plist" | grep Usage
# should list NSMicrophoneUsageDescription, NSSpeechRecognitionUsageDescription, NSAppleEventsUsageDescription

2. Send your friend the .dmg + these steps

Copy-paste this to them:

  1. Open the Echo .dmg and drag Echo into Applications.
  2. Try to open Echo. macOS will block it (it's not from the App Store).
    • If it says "Echo is damaged", open Terminal and paste: 
      xattr -dr com.apple.quarantine /Applications/Echo.app
      then open Echo again.
    • Otherwise: System Settings → Privacy & Security, scroll down, click "Open Anyway" next to Echo.
  3. On first launch, grant the permissions Echo asks for:
    • Microphone (to hear you)
    • Accessibility (to type the text for you)
    • Input Monitoring (for the fn hotkey) — then quit and reopen Echo
  4. On the Transcription step, click Setup Whisper. This downloads the speech model (~142 MB) once — you'll need internet for this. After it finishes, transcription works fully offline.
  5. On the Refinement step, get a free key at console.groq.com and paste it in. This is what cleans up grammar and spelling. (Skip it and you'll still get raw transcription, just without the polish.) Transcription itself works offline with no key.
  6. Press fn (or the menu-bar icon) and start talking. That's it.

Why the Gatekeeper step exists

The build isn't Apple-notarized (that needs a $99/yr Apple Developer account). The xattr / "Open Anyway" step is the standard one-time way to run an unsigned app. To remove it entirely later, enroll in the Apple Developer Program and add signing + notarization secrets to .github/workflows/release.yml.