Transcribe a recorded lecture or meeting into searchable text without uploading it anywhere.
Generate SRT or WebVTT subtitle files for a video using local speech recognition.
Dictate notes through your microphone and have the app type them out in real time.
Translate a transcribed audio file to another language completely offline.
| maddylaneeee/shengji | akshaymaloo/filament | crafcat7/peakmon | |
|---|---|---|---|
| Stars | 7 | 7 | 7 |
| Language | Swift | Swift | Swift |
| Setup difficulty | moderate | easy | easy |
| Complexity | 2/5 | 2/5 | 3/5 |
| Audience | general | general | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires macOS 15.5 on an Apple silicon Mac, and the first launch needs a manual security override because the app is not notarized by Apple.
ShengJi, also called LocalScribe, is a free and open source macOS application that turns audio into text and subtitles on your own machine. It works with microphone input, audio and video files, and audio playing directly from your Mac. The app does not require an account or an internet connection to do its core work, so your audio stays private. It requires macOS 15.5 or later and an Apple silicon Mac. What sets ShengJi apart is its choice of speech recognition engines. Instead of locking you into one system, the app offers Apple Speech, whisper.cpp with Metal acceleration, SenseVoice, and NVIDIA Parakeet. Apple Speech handles microphones, files, and live captions, while the others currently focus on file transcription. The app also includes optional offline translation through Apple Translation and NLLB, so you can convert speech to text and then translate it without sending anything to a remote server. The app is built as a long-task tool. It supports progressive results as audio is processed, pause and resume for recordings, and session recovery journals so you do not lose work on long sessions. Transcripts are fully editable with search, find and replace, trimming, and deletion. You can import and export in several formats, including TXT, Markdown, JSON, PDF, SRT, and WebVTT, which covers most subtitle and document needs. The interface is written in SwiftUI and switches automatically between English and Simplified Chinese based on your Mac language preference. The localization system is designed so that adding another language does not require changes to the feature code. A command-line interface is also available for running models, transcription, export, and translation from the terminal. The project is actively developed but has some limitations. Apple SpeechAnalyzer and floating live captions require macOS 26, while older versions fall back to the other engines. The release build is ad-hoc signed rather than notarized by Apple, so first launch requires a manual override in System Settings. The full README is longer than what was shown.
A free macOS app that turns audio and video into text and subtitles entirely on your own machine. It keeps your data private by working offline without requiring an account.
Mainly Swift. The stack also includes Swift, SwiftUI, whisper.cpp.
The explanation does not mention a specific software license, so the exact permissions are unknown.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly general.
This repo across BitVibe Labs
Verify against the repo before relying on details.