Fast, private, on-device dictation for Apple silicon Macs.

LocalFlow turns speech into polished text at the active cursor without sending audio or transcripts to a cloud service. Hold ⌥ Option + Space, speak, and release. Apple SpeechTranscriber handles transcription and Apple Foundation Models can clean the result before LocalFlow pastes it into the focused app.

What it does

• Streams speech recognition locally with Apple SpeechTranscriber.
• Uses Apple Foundation Models for optional two-second smart cleanup.
• Falls back to the raw transcript if cleanup misses its deadline.
• Pastes safely into the currently focused text field.
• Keeps listening and processing in a 170 × 38 waveform-only notch. Dictation auto-pastes into a focused text field; without an editable target, the notch expands with Copy and History.
• Keeps unlimited, searchable transcript history in a user-private local SQLite database.
• Supports click-to-start/stop, push-to-talk, Reduce Motion, and a persistent 220–420 pt expanded-HUD width.
• Never persists audio, prompts, or clipboard contents. Transcript history is stored only on your Mac and can be cleared at any time.

Requirements

• Apple silicon Mac (MacBook Air, MacBook Pro, Mac mini, Mac Studio, or iMac)
• macOS 26.0 or newer
• Xcode 26 or the matching Swift 6.2 command-line tools
• Microphone, Speech Recognition, and Accessibility permissions

Install

Start with SETUP.md. On a new Mac, the guided flow is:

git clone https://github.com/codejunkie99/LocalFlow.git
cd LocalFlow
./scripts/setup-localflow.sh

This installs one canonical signed copy at ~/Applications/LocalFlow.app and preserves permissions across updates with a stable per-Mac signing identity.

Quick start

git clone https://github.com/codejunkie99/LocalFlow.git
cd LocalFlow
./scripts/package-app.sh
open dist/LocalFlow.app

For a full first-run walkthrough, see Installation.

Use

1. Focus a text field in Notes, TextEdit, Safari, Slack, or your editor.
2. Hold ⌥ Option + Space.
3. Speak naturally.
4. Release the keys to transcribe, optionally clean up, and paste.

A quick tap of the shortcut enables toggle mode: tap once to start and again to stop. You can also use Start Dictation from the LocalFlow menu-bar menu.

Privacy model

• No transcript content is uploaded or logged.
• Audio exists only in memory while recording.
• Transcript history is stored locally at ~/Library/Application Support/LocalFlow/history.sqlite3 for the current user only.
• Use Clear Transcript History in LocalFlow to permanently delete saved transcripts; permissions and settings are unchanged.
• Cleanup prompts and clipboard snapshots are never persisted.
• Logs contain state and timing information only.
• Clipboard contents are restored after the paste attempt when possible.
• Apple may download system-managed speech and language-model assets.

For first-run setup on a new Mac, start at SETUP.md. See Architecture for the trust boundaries and data flow.

Update

Use Update LocalFlow from the LocalFlow menu. It checks the newest tagged source release only when you ask, verifies the manifest/checksum/archive, builds with the same per-Mac signing identity, backs up history, and replaces ~/Applications/LocalFlow.app with rollback safety. There is no background updater. Maintainers publish source releases with ./scripts/publish-source-release.sh; a notarized DMG is a later distribution layer for when Developer ID credentials exist.

Build and test

swift build --disable-sandbox
./scripts/run-tests.sh

Create a release application bundle:

./scripts/package-app.sh

Use a specific signing identity when needed:

LOCALFLOW_SIGNING_IDENTITY="Apple Development: Your Name (TEAMID)" ./scripts/package-app.sh

Stable signing lets macOS keep the Accessibility grant across rebuilds. An ad-hoc signature is fine for a quick local build, but macOS may ask for Accessibility again after the executable changes.

Performance target

For 20 warm trials:

• Median total latency: ≤ 1,500 ms
• p95 total latency: ≤ 2,500 ms
• Paste latency: ≤ 100 ms

Use artifacts/performance/acceptance-template.json with scripts/acceptance-stats.sh to record a local run.

Project layout

Sources/LocalFlowApp       SwiftUI menu-bar app and notch HUD
Sources/LocalFlowCore      State machine, deadlines, and privacy-safe policies
Sources/LocalFlowPlatform  Apple Speech, Foundation Models, shortcut, and paste adapters
Tests                      Deterministic core and platform checks
scripts                    Build, package, test, and acceptance helpers

Contributing

Bug reports and focused pull requests are welcome. Read CONTRIBUTING.md before making changes. Security and privacy issues should follow SECURITY.md.

License

LocalFlow is available under the MIT License.