How it works

From page to voice, on your device

PageNarrator finds the useful text, turns it into speech on your device, and keeps the words on screen in step with the audio.

1. You choose what to read

Select a passage or ask PageNarrator to read the whole page. It looks for the main content and leaves out much of the navigation, advertising, and other clutter.

2. The voice runs on your device

The built-in Kokoro voices run on your device. When you prepare a voice or start reading, missing voice files download from Hugging Face. With a compatible GPU, PageNarrator uses WebGPU for faster speech. Without one, it runs on your CPU. After a successful download, the same voice and processing mode can reuse cached files offline; clearing the cache or changing mode may require another download.

3. The page follows the narration

PageNarrator builds a transcript, highlights the current passage, and scrolls it into view. You can pause, jump between sections, change speed, and return to a saved position.

4. PDFs stay readable

PDFs open in PageNarrator’s own viewer, which extracts the text and highlights it as it is read. The document stays in view, and the controls stay in the toolbar.

Learn about PDF reading.