Velora 0.1.0
First public build of Velora — local-first, on-device dictation for macOS (Apple Silicon).
Install
- Download
Velora-0.1.0.dmgbelow, open it, drag Velora to Applications. - This build is self-signed, not yet notarized, so the first open needs right-click → Open (or System Settings → Privacy & Security → Open Anyway).
- On first launch, grant Microphone, Input Monitoring, and Accessibility when prompted — all three are required for the hotkey and text insertion to work.
- The engine downloads its models (~4 GB) from Hugging Face on first run; after that everything runs offline.
Prefer to build from source? See the README — make app.
Highlights
- Multilingual dictation (English, Indian English, Hindi + more) via on-device
whisper-large-v3-turbo - App-aware LLM cleanup with custom modes/prompts
- History browser, 6-month audio archive + reprocess, optional romanized (Hinglish) output
- Zero network calls at dictation time
⚠️ Requires Apple Silicon (M1+) and macOS 14+.