Hotfix

Fixed AI output being silently truncated with certain model providers.

Number Conversion

Spoken or typed number words are now automatically converted to numerals — so "twenty-three" becomes "23" and "one hundred" becomes "100".

Bug Fixes

Fixed compatibility issues with local Ollama endpoints for AI post-processing.

Dedicated Dictionary Experience

A much stronger dictionary experience with a dedicated dictionary page, searchable entries, word cloud visualization, and easier management of learned and corrected terms.

Offline Dictionary Corrections

Improved the offline dictionary correction workflow so local corrections are applied more reliably and feed better into future transcription behavior.

Smarter Adaptive Learning

Upgraded the adaptive dictionary learning logic to detect real corrections more accurately and learn more cautiously, reducing bad auto-learning.

Cleaner AI Output

Improved AI prompt quality and post-processing rules to produce cleaner, more consistent final text with better rewriting behavior.

Transcription Cleanup

Strengthened transcription cleanup to reduce filler and repeated-word artifacts, improving readability of final output.

Faster Live Dictation

Refined overall transcription pipeline performance for faster, smoother live dictation startup and better end-to-end responsiveness.

Context Gathering & Export

Much better context gathering for more accurate transcriptions, plus the ability to export transcriptions, clips, and board items.

Onboarding & Settings

Updated onboarding and settings flows to make feature controls clearer and reduce friction for existing users.

Text-to-Speech Integration

A brand new, high-quality TTS engine that delivers professional-grade speech synthesis. Seamlessly convert your transcriptions back to speech with natural, fluent voice output.

Dock & Menu Bar Control Restored

Welcome back the ability to show or hide the app icon from the Dock. Manage your menu bar clutter with full control over Pipit's presence.

Improved Keychain Handling

More reliable and graceful handling of AI API keys stored in the system keychain.

Bug Fixes

• AI Enhance hotkey now properly deactivates when AI processing is disabled
• Model downloads now persist reliably between app launches, eliminating redundant downloads
• Minor startup performance improvements for a snappier experience

Sound Effects Fixed

Audio feedback now plays correctly during transcription.

Auto-Update Fixed

Sparkle auto-update now downloads and installs correctly. Improved update window visibility for menu bar apps.

Version Detection

Enhanced version comparison for more dependable update detection.

Bug Fixes

Various stability improvements and minor bug fixes.

Menu Bar Visibility

Improved menu bar icon visibility when the app window is open.

Bug Fixes

Various stability improvements and minor bug fixes.

Voice Transcription

System-wide voice-to-text with local processing. Works in any app via global hotkeys.

Unified History

Clipboard and transcription history in one searchable menu bar interface.

Privacy First

100% local processing. No cloud required, no data transmission. Your voice stays on your Mac.

Language Support

Optimized English mode or multi-language support for 25+ European languages.

AI Post-Processing

Optional AI enhancement with your own API key for grammar and formatting improvements.

File Upload & Transcribe

Drop audio files for instant transcription of meetings, interviews, and voice memos.

Performance Optimized

Low-latency transcription, battery-optimized modes, instant text injection.

Adaptive Dictionary

Learns your vocabulary over time. Names, jargon, technical terms—automatically corrects common mishearings.

Quick Actions

Voice commands for your Mac. "Open Spotify", "Search Google for...", "Ask Claude..."—hands-free control.