1.1.166341.1 MB
Proprietary
strict
core24
Voice-to-text for Linux. One hotkey, any app. AI cleans your speech into polished text. E2E encrypted, privacy-first.
AICHE is a voice-to-text tool for Linux that works in any application. Press one hotkey, speak naturally, and get clean, polished text pasted directly where your cursor is - in your terminal, editor, browser, chat app, or anything else that accepts text input.
Works with 100+ applications including Cursor, VS Code, Slack, Discord, Gmail, Notion, GitHub, Linear, ChatGPT, and Claude. If you can type into it, you can dictate into it.
Universal hotkey. One key works in every app, system-wide. No copy-paste workflow, no per-app configuration. Press, speak, release - text appears where you're typing.
Voice Code mode. Hands-free dictation for developers with neural voice activity detection. Voice commands for navigating Cursor and VS Code numbered suggestions, dictating code comments, and controlling your editor without touching the keyboard.
Sub-3-second transcription. 15 minutes of audio processed in about 3 seconds on average. Long recordings don't mean long waits.
AI cleanup, not raw transcription. Filler words removed automatically. Grammar corrected. Punctuation added. The output is text you'd actually send, not a raw speech dump you have to clean up afterwards.
Personal Dictionary. Teach the AI your names, acronyms, codebase terms, and domain jargon once - it remembers them across every device and every transcription.
Cross-device sync. Start dictating on Linux, continue on your phone, review on your tablet. Same account, same history, same dictionary everywhere. Desktop clients for Linux, Mac, and Windows. Mobile apps for iOS and Android. Apple Watch app. Chrome extension. REST API.
Privacy
End-to-end encrypted cloud sync using AES-256-GCM with Argon2id key derivation. Your passphrase never leaves your device - the server cannot read your data, even if compelled. Audio files are purged within 15 seconds of transcription. Zero long-term retention on servers. No telemetry, no tracking, no training data collection from your recordings.
For developers
Voice Code mode was built for the AI-assisted coding workflow. Dictate prompts into Cursor, Claude Code, or ChatGPT without breaking flow. Speak commit messages instead of typing them. Voice-navigate Copilot suggestions. Pair-program with your AI tools at conversation speed instead of typing speed.
Offline and resilience
Recordings queue locally when offline and sync when the connection returns. Nothing is lost if your network drops mid-sentence.
Auto-Send
Optional mode that pastes your transcription and presses Enter automatically - useful for chat apps, terminal commands, and form submissions where you don't want the extra keystroke.
Free trial
7-day free trial with full access. No credit card required.
Links
Works with 100+ applications including Cursor, VS Code, Slack, Discord, Gmail, Notion, GitHub, Linear, ChatGPT, and Claude. If you can type into it, you can dictate into it.
Universal hotkey. One key works in every app, system-wide. No copy-paste workflow, no per-app configuration. Press, speak, release - text appears where you're typing.
Voice Code mode. Hands-free dictation for developers with neural voice activity detection. Voice commands for navigating Cursor and VS Code numbered suggestions, dictating code comments, and controlling your editor without touching the keyboard.
Sub-3-second transcription. 15 minutes of audio processed in about 3 seconds on average. Long recordings don't mean long waits.
AI cleanup, not raw transcription. Filler words removed automatically. Grammar corrected. Punctuation added. The output is text you'd actually send, not a raw speech dump you have to clean up afterwards.
Personal Dictionary. Teach the AI your names, acronyms, codebase terms, and domain jargon once - it remembers them across every device and every transcription.
Cross-device sync. Start dictating on Linux, continue on your phone, review on your tablet. Same account, same history, same dictionary everywhere. Desktop clients for Linux, Mac, and Windows. Mobile apps for iOS and Android. Apple Watch app. Chrome extension. REST API.
Privacy
End-to-end encrypted cloud sync using AES-256-GCM with Argon2id key derivation. Your passphrase never leaves your device - the server cannot read your data, even if compelled. Audio files are purged within 15 seconds of transcription. Zero long-term retention on servers. No telemetry, no tracking, no training data collection from your recordings.
For developers
Voice Code mode was built for the AI-assisted coding workflow. Dictate prompts into Cursor, Claude Code, or ChatGPT without breaking flow. Speak commit messages instead of typing them. Voice-navigate Copilot suggestions. Pair-program with your AI tools at conversation speed instead of typing speed.
Offline and resilience
Recordings queue locally when offline and sync when the connection returns. Nothing is lost if your network drops mid-sentence.
Auto-Send
Optional mode that pastes your transcription and presses Enter automatically - useful for chat apps, terminal commands, and form submissions where you don't want the extra keystroke.
Free trial
7-day free trial with full access. No credit card required.
Links
- Homepage: https://aiche.app
- Documentation: https://aiche.app/docs
- Privacy policy: https://aiche.app/privacy
- Terms of use: https://aiche.app/terms
Update History
1.1.16 5 → 622 Jul 2026, 17:00 UTC
1.1.2 (3) → 1.1.16 (5)22 Jul 2026, 13:00 UTC
1.1.2 (3)16 Jul 2026, 04:15 UTC
16 Jul 2026, 04:03 UTC
22 Jul 2026, 15:57 UTC
16 Jul 2026, 04:15 UTC