Transcription (using Audio Prompt API)

This page contains an example of Transcription using the Audio Prompt API.


Make sure that all the requirements are green.

Activate chrome://flags/#chunkInterval-api-for-gemini-nano-multimodal-input
Loading...
Pending

Activate all the flags
FlagValue to select
chrome://flags/#optimization-guide-on-device-modelEnabled Bypass*
chrome://flags/#prompt-api-for-gemini-nano-multimodal-inputEnabled
chrome://flags/#optimization-guide-debug-logsEnabled
Verify that Gemini Nano has been downloaded
  1. Navigate to chrome://components
  2. Search for Optimization Guide On Device Model on the page.
  3. Click on Check for update
  4. Ensure that it says Up-to-date

Audio is limited to 30 secondes

Audio support right now is limited to 30 seconds. However, see the Long Audio section for how to circumvent this limitations.

It seems like the page is stuck

After you have clicked on Execute for the first time, Chrome has to download Gemini Nano. This can take a while (5-10 minutes). While this is happening, the page can seem frozen or like nothing is happening. Give it time then try to restart your browser.

Playground