v139 · Web APIs · Speech

On-device Web Speech API

Chrome 139 starts the behind-a-flag developer trial for on-device speech recognition in the Web Speech API. Pages can ask for local processing with the same familiar SpeechRecognition interface, then check and install language packs before starting recognition.

concepts

  1. Feature Detection & Setup

    Checks whether on-device speech recognition is available in this browser, shows the correct configuration to request local processing, and provides a live transcription widget with microphone access.

  2. Privacy Architecture Comparison

    Side-by-side diagram of what data flows to the cloud in classic Web Speech vs what stays entirely on the device with on-device recognition — highlighting the privacy implications for each approach.

  3. Availability & Language Matrix

    Calls SpeechRecognition.available({ langs, processLocally, quality }) for a dozen locales and compares local-only against remote-allowed recognition. Click "install" to trigger SpeechRecognition.install(); click "start recognition" to run a single-utterance test with processLocally = true.

  4. Offline Dictation Pad

    Continuous dictation using processLocally: true. Interim results appear in italic below the textarea; final results append above. A network status badge responds to real online/offline events, showing that on-device recognition works in both states. Stats panel tracks word count, confidence, and processing mode.

why it shipped

The Web Speech API has been in Chrome since version 25, but recognition could be handled by a cloud service, sending audio away from the page and depending on a network round trip. The v139 developer trial adds an explicit local-processing path: developers check whether a language pack is available, install it when needed, and set processLocally before recognition starts. That makes the privacy and offline guarantees visible to the application instead of leaving processing location implicit.

references