Feeding a clanker? Grab this page as raw .md →

WebkitSpeechToTextWidget API#

WebkitSpeechToTextWidget is a thin bridge to the browser's WebKit speech recognition API: press the button, talk, and the recognized text lands in transcript as a Python string. Setting trigger_listen starts a session from code instead of the button, and listening reflects whether recognition is running. Because it delegates to the browser, availability depends on the browser — WebKit speech recognition is a Chrome/Safari feature and is absent or unreliable elsewhere, so treat it as best-effort rather than a dependable transcription pipeline.

See also: AnnotationWidget for labelling with speech, keys and buttons in one surface, KeystrokeWidget for keyboard input, and WebcamCapture for pulling video frames from the same browser APIs.

Bases: AnyWidget

Speech-to-text widget backed by the browser's Webkit Speech API.

The widget exposes the transcript text along with the listening and trigger_listen booleans; it does not require initialization arguments.

Examples:

import marimo as mo
from wigglystuff import WebkitSpeechToTextWidget

speech = mo.ui.anywidget(WebkitSpeechToTextWidget())
speech

Synced traitlets#

Traitlet	Type	Notes
`transcript`	`str`	Latest transcript from the browser.
`listening`	`bool`	Whether speech recognition is active.
`trigger_listen`	`bool`	Toggle listening when set to true (auto-resets).