Your voice, unleashed.
What if you could type into any app, just by speaking —
IfVoice is an AI voice input interface for macOS. Press a hotkey, speak, and your words become text.
Everything stays local.
Your own speech recognition engine, independent of the cloud.
Fully Local Processing
High-speed inference via Metal / CoreML. No internet required — your voice data never leaves your device.
Context-Aware Pro
Formal in Mail, casual in Slack, code in VS Code. Text style adapts automatically to the active app.
Sub-Second Latency
Text appears the moment you release the hotkey. Real-time streaming recognition with zero wait.
Auto Meeting Minutes
Dual recording of mic + system audio with speaker diarization, summaries, and Markdown export.
12 Languages
High-accuracy recognition for Japanese, English, Chinese, Korean, French, and 7 more languages.
Custom Hotkeys
Trigger with your preferred key combination. Supports both push-to-talk and toggle recording.
Simple as 1-2-3
Install and start speaking immediately.
Press the Hotkey
Press your configured hotkey in any app to start recording.
Speak
Just talk into your microphone. Speech is recognized in real-time.
Text Auto-Inserts
Release the hotkey and formatted text is typed at the cursor position.
Different app, different voice.
The same speech is formatted optimally for each context.
"The meeting tomorrow has been moved to 3 PM. Please notify all participants."
Dear team,
Please be advised that tomorrow's meeting has been rescheduled to 3:00 PM.
Kindly notify all participants at your earliest convenience.
Best regards
// TODO: Meeting rescheduled to 15:00 tomorrow // - Notify all participants
The meeting tomorrow has been moved to 3 PM. Please notify all participants.
Your voice belongs to you.
IfVoice is designed privacy-first from the ground up.
Local by Default
Speech recognition runs entirely on-device. No data is ever sent to external servers.
Cloud is Opt-In
Cloud API usage is completely optional. Use your own API keys with your preferred service.
Transparency Promise
Open source. All code is publicly available for anyone to audit at any time.
Free your meetings from note-taking.
Dual recording and AI summaries automate your meeting minutes.
Dual Recording
Captures both mic and system audio simultaneously
Speaker Diarization
Automatically identifies and labels speakers
Markdown Export
Outputs structured minutes instantly
AI Summary
Auto-summarizes key meeting points
Find your plan.
Both plans include a 7-day free trial. Try before you decide.
Pro One-Time
All features, pay once
- Local speech recognition (12 languages)
- Custom hotkeys
- LLM text formatting (BYOK)
- Context-aware text formatting
- Unlimited meetings (90 min)
- Meeting summaries & export
- Large STT models
No update guarantee after 6 months
Try free for 7 daysPro Subscription
Always up-to-date + future cloud features
- Local speech recognition (12 languages)
- Custom hotkeys
- LLM text formatting (BYOK)
- Context-aware text formatting
- Unlimited meetings (90 min)
- Meeting summaries & export
- Large STT models
- Update guarantee
Cloud integration planned for future releases
Try free for 7 days