Scribis: Speech to Text

Scribis is a high-performance local AI speech-to-text tool tailored for the Apple Silicon chip series. We do more than just convert speech to text; powered by a robust local AI engine, we transform complex audio tasks into ultimate productivity. All processing is completed entirely on your device, ensuring absolute privacy and zero server costs.

Supports Top-Tier Local AI Models:

* OpenAI Whisper: Globally recognized as the most powerful general-purpose speech recognition model.
* Advanced Engines: Built-in support for SenseVoice Small, Parakeet, Qwen 3 ASR, VibeVoice, and Voxtral, balancing speed and extreme precision.

Precision Video OCR & Text Extraction:

* Burnt-In Subtitle Extraction: Accurately extract difficult hardcoded video subtitles. Use the custom zone framing tool to target text and instantly digitize it into a transcript.

Real-Time Transcription & Smart Audio Capture:

* App-Specific Audio Capture: Select exactly which application's system audio to record, filtering out unwanted background noise.
* Floating Subtitle Widget: Record live audio and view instant transcripts in a fully movable, always-on-top floating window.
* Global Translation: Break language barriers in real-time during international conferences or live foreign-language lectures.

Smart Speaker Diarization:

* Automatic Role Separation: Accurately distinguish between different speakers in the audio.
* The Ultimate Meeting Tool: Automatically label who said what and when, keeping transcripts organized and clear at a glance.

AI Smart Assistant & Local LLM Support:
More than just transcripts, Scribis is your second offline brain:

* Local GGUF Integration: Import and run custom offline language models (GGUF format), ensuring complete data privacy.
* One-Click Summary & Chat: Generate refined core summaries for hours of meetings in seconds, and interact directly with your audio content to retrieve missed details.

Text-to-Speech (TTS) & Audio Resynthesis:

* CPU-Powered Voice Conversion: Redesign audio completely offline using the current speaker profile.
* Subtitle-Driven Generation: Reconstruct missing or edited speech directly from your subtitle text.
* Native TTS Engine: Currently supports English, Chinese, and Japanese with highly realistic voices. More languages will be supported in the future as we continue to train custom models.

Professional Workflow Integration:

* Multi-Format Export: Supports SRT, VTT, Markdown, and plain text export, integrating seamlessly with editing software.
* Efficient Media Library: Easily manage thousands of transcription projects with fast searching and categorization.

Supported Transcription Languages:
Traditional/Simplified Chinese, English, Japanese, Korean, Cantonese, Vietnamese, French, German, Spanish, Italian, Portuguese, Russian, Thai, Hindi, Arabic... and over 99 other languages.

Who is Scribis for?

* Video Creators: Quickly extract hardcoded text, generate high-precision subtitles, and enter the global market.
* Business Professionals: Transform multi-person meetings into organized action guides completely offline.
* Journalists & Interviewers: Quickly distinguish between speakers and instantly summarize key points.
* Students & Researchers: Automatically transform classroom recordings into structured notes.

Download Scribis now and start your new era of local intelligent productivity.

Terms of Use (EULA): https://www.apple.com/legal/internet-services/itunes/dev/stdeula/
Privacy Policy: https://www.scribis.app/en/privacy/