This release includes 2 breaking changes for platform teams planning a safe upgrade.
✓ No known CVEs patched in this version
Topics
+5 more
ReleasePort's take
Moderate signalTranscription now operates fully on‑device via Apple's SpeechAnalyzer; the cloud transcription stack has been removed. The project name changed from FreeFlow to Megaphone.
Why it matters: All transcription processing moves locally, eliminating reliance on remote services and reducing latency for users with unstable connections. Renaming impacts branding, documentation, and any integration points referencing "FreeFlow".
Summary
AI summaryTranscription runs entirely on‑device via Apple's SpeechAnalyzer, requiring macOS 26.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Security | Medium |
App notarization missing; macOS warns of potential malware risk. App notarization missing; macOS warns of potential malware risk. Source: granite4.1:30b@2026-07-17-audit Confidence: low |
— |
| Breaking | High |
Transcription runs entirely on‑device using Apple's SpeechAnalyzer; cloud transcription stack removed. Transcription runs entirely on‑device using Apple's SpeechAnalyzer; cloud transcription stack removed. Source: llm_adapter@2026-07-17 Confidence: high |
— |
| Breaking | High |
Renamed project from FreeFlow to Megaphone. Renamed project from FreeFlow to Megaphone. Source: granite4.1:30b@2026-07-17-audit Confidence: low |
— |
| Feature | Medium |
Recording start, stop, and error feedback sounds are configurable in Settings with preview. Recording start, stop, and error feedback sounds are configurable in Settings with preview. Source: llm_adapter@2026-07-17 Confidence: high |
— |
| Feature | Medium |
Custom vocabulary now biases the on‑device speech model directly via analyzer contextual strings. Custom vocabulary now biases the on‑device speech model directly via analyzer contextual strings. Source: llm_adapter@2026-07-17 Confidence: high |
— |
| Dependency | High |
Requires macOS 26 (Tahoe) on Apple silicon hardware. Requires macOS 26 (Tahoe) on Apple silicon hardware. Source: granite4.1:30b@2026-07-17-audit Confidence: low |
— |
| Bugfix | Low |
Transcription language picker now lists languages supported by the on‑device model; Settings shows install status with download button. Transcription language picker now lists languages supported by the on‑device model; Settings shows install status with download button. Source: llm_adapter@2026-07-17 Confidence: low |
— |
Full changelog
[1.0.0] - 2026-07-14
First Megaphone release, forked from FreeFlow 1.1.0.
Changed
- Renamed to Megaphone, with a new liquid-glass monochrome megaphone icon.
- Transcription now runs entirely on-device via Apple's SpeechAnalyzer (requires macOS 26). Audio is analyzed while you speak, so the transcript is ready almost instantly when you stop — and recordings never leave your Mac. The cloud transcription stack (transcription provider/model settings, transcription API URL and key, realtime WebSocket streaming, and background HTTP pre-transcription) has been removed; the configured API key is now used only for LLM cleanup and app context.
- The transcription language picker now lists the languages supported by the on-device speech model, and Settings shows the model's install status with a download button.
- Custom vocabulary now also biases the on-device speech model directly (via analyzer contextual strings), in addition to guiding LLM cleanup.
Added
- The recording start, stop, and error feedback sounds are now configurable in Settings, with a picker and preview button for each event across the full set of built-in macOS alert sounds.
Download
Megaphone.dmg — requires macOS 26 (Tahoe) on Apple silicon.
[!NOTE]
This build is not yet notarized by Apple, so macOS will warn it "could not verify Megaphone.dmg is free of malware." Runxattr -d com.apple.quarantine ~/Downloads/Megaphone.dmgbefore opening, or use System Settings → Privacy & Security → Open Anyway.
Breaking Changes
- Removed cloud transcription stack (API URL, key, WebSocket streaming, background HTTP pre‑transcription).
- Transcription now requires macOS 26 due to on‑device Apple SpeechAnalyzer usage.
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About Megaphone
All releases →Related context
Related tools
Beta — feedback welcome: [email protected]