Release Notes
Version history for Flow Keyboard.
v0.2.7
March 25, 2026
- Bug fixes and improvements
v0.2.6
March 21, 2026
- Bug fixes and improvements
v0.2.5
March 20, 2026
- Editing mode now supports multi-step undo and redo
- Added line break and hyphen to the punctuation popup
- Bug fixes and improvements
v0.2.4
March 13, 2026
New Features
- Built-in QWERTY keyboard with letters, numbers & symbols — no need to switch input methods
- Add to Dictionary — long-press to select text in any app and save it to your personal dictionary
Improvements
- Auto-close timer now defaults to 5 minutes, with new 5-second and 1-minute options to reduce microphone indicator display time
- Faster voice recognition with lower latency
- Other fixes and improvements
Tip: After speech-to-text finishes, tap the Edit button in the bottom-right corner to refine the result — just say what you'd like to change.
v0.2.3
March 5, 2026
First Public ReleaseSpeak your thoughts naturally and get polished, professional text — no need to draft sentences or worry about grammar. Flow Keyboard is an AI voice typing keyboard that turns your speech into clear, structured writing in seconds.
AI-Powered Voice to Text
- Natural speech recognition that handles pauses, filler words, and incomplete sentences
- Automatic removal of verbal tics like "umm" and "uh"
- Built-in grammar correction, punctuation, and logical restructuring
Advanced Refinement
- Combines transcription with AI editing in a single step — processes messy spoken input into polished text without manual cleanup
Multilingual Support
- 30+ languages with automatic language detection
- Mixed Chinese/English recognition
- Translation mode — speak in one language, output in another (16 target languages)
System-Wide Keyboard
- Works across Messages, Mail, Notes, WeChat, Slack, and all other apps
- Requires iOS 17.0 or later
- 6.5 MB install size
Use Cases
- Quick message and email replies
- Report writing and meeting notes
- Brainstorming and voice-assisted coding prompts
Privacy
- Audio is processed and immediately discarded — zero server-side audio retention
v0.7.10
March 25, 2026
Public Beta- Bug fixes and improvements
v0.7.9
March 19, 2026
Public Beta- Bug fixes and improvements
v0.7.8
March 14, 2026
Public Beta- Bug fixes and improvements
v0.7.7
March 2, 2026
Public Beta- Bug fixes and improvements
v0.7.5
February 25, 2026
Public Beta- Bug fixes and improvements
v0.7.4
February 20, 2026
Public Beta- Bug fixes and improvements
v0.7.2
February 15, 2026
Public BetaAI Voice Input
- Speak naturally and get polished, well-structured text -- filler words removed, grammar fixed, punctuation added automatically
- 30+ language recognition with automatic mixed-language detection -- no manual switching needed
- Translation mode -- speak in one language, output text in another (16 target languages)
- Under 2 seconds for short recordings, up to 5 minutes maximum
Smart Output
- Vibe coding mode -- structures speech into clear prompts for AI coding tools (Claude Code, Codex, ChatGPT)
- Professional tone for workplace emails and documents
- Emotional tone preservation in casual chat (Slack, iMessage, Discord)
- Custom dictionary -- add proper nouns, brand names, and technical terms for higher accuracy
Edit Mode
- Select existing text and refine it with voice
Privacy & Performance
- Zero server-side data retention -- all audio and text discarded after processing
- No surrounding text collection -- only your voice is sent to servers
- Native macOS app built with SwiftUI (< 30 MB installer)
- Works system-wide in any macOS app with a text field