LocalScribe — Private Tab Transcription

Turn audio playing in your current Chrome tab into editable text with LocalScribe. LocalScribe captures audio from a tab you explicitly authorize and transcribes it locally on your device using Whisper. Your captured audio and transcript are not uploaded to a transcription service. KEY FEATURES • Private, local transcription Audio processing and speech recognition run on your device. • Capture audio from the current tab Transcribe spoken content playing in supported Chrome tabs. You control when capture starts and stops. • No account or sign-up Start transcribing without creating an account or sending audio to a transcription API. • 99 supported languages Choose a spoken language or let Whisper detect it automatically. • Two accuracy options Use the recommended Fast model for quicker processing or choose Higher Accuracy when additional processing time is acceptable. • Prepare the model before recording Set up the selected speech-recognition model in advance, or let LocalScribe prepare it automatically while capturing audio. Model files are saved by Chrome for later use when available. • Editable transcript Review and edit the generated text directly in the side panel. • Timestamped transcript view Switch between a clean editable transcript and a read-only view showing the original Whisper segment timestamps. • Subtitle and text exports Download your result as: – TXT for plain text – SRT for subtitles – VTT for web captions • Persistent side-panel workflow Continue viewing the webpage while LocalScribe captures and processes its audio. • Four interface languages The LocalScribe interface is available in English, Spanish, Portuguese, and French. HOW IT WORKS 1. Open a normal webpage containing audio or video. 2. Click the LocalScribe toolbar icon on that tab. 3. Select the spoken language and accuracy option. 4. Optionally prepare the selected model before recording. 5. Start capturing the tab. 6. Stop when the section you need has finished. 7. Review, edit, copy, or download the transcript. PRIVACY LocalScribe is designed for private, on-device transcription: • Captured audio is processed locally. • Transcripts remain on your device unless you choose to copy or download them. • No LocalScribe account is required. • No transcription API is used. • No analytics or advertising SDK is included. The speech-recognition model files are downloaded from the model provider during initial setup and stored by Chrome when possible. An internet connection is therefore required for first-time model preparation or if Chrome has removed previously stored model data. PERMISSIONS LocalScribe requests only the browser capabilities needed for its features: • Active tab and tab capture: capture audio only from the tab you explicitly authorize by clicking the extension icon. • Side panel: display transcription controls and results beside the webpage. • Offscreen document: maintain authorized audio capture and perform local processing. • Storage: preserve extension state, model readiness information, and operational diagnostics. LocalScribe does not automatically capture tabs in the background. Capture begins only after you open the extension from a webpage and select the start action. IMPORTANT LIMITATIONS • Chrome internal pages cannot be captured. • Protected or DRM-controlled media may prevent audio capture. • Initial model preparation may take time and use significant bandwidth. • Local transcription speed depends on the device, recording duration, and selected accuracy option. • Transcripts and timestamps are generated automatically and may contain errors. Review important content before publishing or relying on it. • LocalScribe does not identify people by name. Whisper is an open-source speech-recognition model created by OpenAI. LocalScribe is an independent product and is not affiliated with, endorsed by, or sponsored by OpenAI.