Documentation
Everything you need to use Neuron Flame well — from your first transcript to per-app Power Mode and custom local models.
Quick Start
Install, grant permissions, set a hotkey, transcribe your first thought — all in 60 seconds.
Permissions
Why Neuron Flame asks for Microphone and Accessibility, and how to grant them on macOS.
AI Enhancement
Polish raw transcripts with Gemini, OpenAI, Anthropic, Ollama, or any local CLI.
Power Mode
Auto-swap settings per app — different voice for Slack, Mail, your code editor.
Hotkeys
Push-to-talk, hands-free toggle, paste-only mode, and chord shortcuts.
Custom Models
Compare Whisper and Parakeet, install community models, point at any OpenAI-compatible endpoint.
Privacy
Where your data lives, what leaves your Mac, and exactly when.
Troubleshooting
Hotkey not firing? No paste? Permissions reset after rebuild? Solutions here.
New to Neuron Flame?
Start with Quick Start, then read Permissions. Most people are productive within a minute. Once you're comfortable with push-to-talk, explore Power Mode — it's the feature you didn't know you needed.
Coming from another dictation app?
If you're switching from Apple's built-in dictation or another voice-to-text tool, Neuron Flame's hotkey-driven flow will feel familiar but faster. The key differences:
- Push-to-talk by default — hold the key while speaking, release when done. No "wake word", no timeout.
- Local-first — Parakeet V2 runs on your Apple Neural Engine in milliseconds. No cloud latency.
- Cursor-aware paste — text appears wherever your blinking caret already was.
- AI enhancement is optional and separate — the raw transcript is always saved, even if enhancement fails or is off.