← Changelog

Web

Every Scribe0 for Web release, in plain language.

Open in browser

  1. 27 June 2026

    Added
    • Upload an audio file directly and have it transcribed, instead of only dictating live.
    Fixed
    • Long enhancements no longer time out. The AI enhance step now runs as a background job, so a long consult is no longer cut off partway through.
    • Large audio uploads no longer fail. Audio is staged in storage rather than being pushed through the request body, which had a size cap.

  2. 12 Apr 2026

    Added
    • Versioned AI result tabs. Each enhance produces its own tab, so you can compare outputs instead of overwriting the previous one.
    • Accept or reject an AI result before it feeds anything downstream.
    • Support for custom WebSocket transcription providers, including WhisperLiveKit.
    Fixed
    • A page load error caused by a hydration mismatch.
    • AI diff formatting stayed visible after accepting the first result.

  3. 7 Dec 2025

    First working release. Realtime cloud dictation in the browser.

    Added
    • Realtime dictation with live streaming transcription (AssemblyAI).
    • AI Enhance to turn raw dictation into a structured clinical note, via OpenRouter.
    • Auto enhance when you stop dictating, if you want it.
    • Editable transcripts with start, stop and resume.
    • Voice commands and a review mode.
    • Accounts and workspaces.
    • A live streaming status indicator and latency readout.