Meetily Community 0.3.0 Release Notes

What's New in Community 0.3.0

Our first release after crossing 10,000 GitHub stars - and the biggest Community update yet. This release delivers the most-requested feature (audio file import) and significant quality-of-life improvements under the hood.

Major New Features

Audio File Import

The most requested feature on GitHub and Discord is here. Drag and drop any audio file into Meetily and get it transcribed locally with full privacy.

Supported formats: MP3, WAV, M4A, OGG, FLAC, MKV, WebM, WMA

Files up to 20 GB supported
Same transcription engines as live meetings (Whisper or Parakeet)
99+ languages supported
All processing stays 100% on your device

Retranscription

Re-process past recordings with a different transcription model for better results. Useful when you want to compare models or improve quality on older recordings.

Beta Feature Gating

New system for previewing experimental features before they ship as stable. Audio import and retranscription are the first features to use this system.

Quality Improvements

Better Transcription Quality

Rewrote the resampling pipeline with parallel processing for large audio files. Handles sample rate conversion more reliably and processes faster on multi-core machines.

Smarter Memory Management

The transcription engine now automatically unloads after batch processing jobs complete, freeing system resources. Fixes a memory leak where the engine stayed loaded in RAM indefinitely.

Language Preference Sync

Fixed a desync issue where the transcription language preference could reset on startup. Language settings are now centralized and persist correctly.

VAD Timestamp Accuracy

Fixed ambiguity in Voice Activity Detection timestamp calculations and model key parsing for more accurate speech segment detection.

Bug Fixes

Fixed audio channel count using decoded data instead of metadata
Fixed import dialog re-initializing on config changes
Replaced unstable floor_char_boundary with stable equivalent
Fixed build breakers, race conditions, and data safety issues identified during code review
Resolved memory leaks in import/retranscribe workflows

Download

macOS (Apple Silicon): Download .dmg(opens in new tab)
Windows: Download .exe(opens in new tab)

New Contributors

@jeremi - Built the audio import feature
@p-s-vishnu - Improved transcription quality and parallel resampling

Full Changelog

View on GitHub(opens in new tab)

Release Highlights