LocalScribe icon
LocalScribe icon

LocalScribe

 Like

Open-source Mac application supporting local transcription of microphone, media, and system audio sources, with editable transcripts, live captions, privacy by design, export in TXT, Markdown, JSON, PDF, SRT, WebVTT, translation, and offline processing. Requires Apple silicon.

ShengJi home screen showing local recognition model selection, transcript recovery, transcript import and offline translation options.

Cost / License

  • Free
  • Open Source (MIT)

Application type

Since

  • Since 2026

Updated

  • Updated Oct 2026

Platforms

  • Mac
0likes
0articles
Save

Features

Properties

Features

Highlights All activities

LocalScribe News & Activities

Recent activities

LocalScribe information

AlternativeTo Categories

Audio & Music, Video & Movies, AI Tools & Services

GitHub repository

  •  23 Stars
  •  2 Forks
  •  0 Open Issues
  •   Updated  
View on GitHub
LocalScribe was added to AlternativeTo by shengjiapp on and this page was last updated . LocalScribe is sometimes referred to as ShengJi. 

No comments or reviews, maybe you want to be first?

Official Links

What is LocalScribe?

LocalScribe (formerly ShengJi) is a free, open-source native transcription app for Apple silicon Macs. It turns microphone input, audio/video files and the sound playing on the Mac into editable text and subtitles, with local recognition, transcript recovery, translation and export.

Choose Apple Speech, Whisper, SenseVoice or NVIDIA Parakeet. Apple Speech and floating live captions require macOS 26; SenseVoice and Parakeet currently support file transcription only. Transcripts can be searched, corrected, trimmed and exported as TXT, Markdown, JSON, PDF, SRT or WebVTT.

New in v1.7.0: enable Save Original Audio before starting a new microphone transcription to keep a local M4A recording. Choose from three recording-quality profiles, then replay and seek through the audio at 0.5×, 1× or 2× to check the transcript. The original M4A can be exported alongside the text or subtitles with the same base filename. Editing or translating the transcript does not alter the recording, and the latest recoverable task retains its audio.

Original-audio saving is off by default for each new task. It is available only for new microphone transcriptions, not media-file transcription, Mac system audio, floating captions or microphone transcription appended to an imported transcript.

Optional on-device Gemma 4 tools proofread or summarize transcripts. Gemma loads only when an AI action is selected, rather than automatically after transcription; its helper exits when the AI task finishes. AI Enhancement is off by default on Macs with 8 GB of memory or less and can be enabled manually. AI output still needs review. Recent updates also improve long Chinese/Japanese/Korean subtitle segmentation and language switching.

The interface supports English, Simplified Chinese and Japanese. No account is required. Recognition audio, imported transcripts and AI processing content are not uploaded by the app. Initial model/language downloads and update checks require network access.

Requires macOS 15.5 or later on Apple silicon. The current release is v1.7.0 build 40 and is not Developer ID signed or notarized by Apple.