← All work
Software & AI

AI Audio Factory

A local application that turns PDFs into chaptered audiobooks.

Audiobook Factory empty PDF upload panel with Kokoro voice and Qwen3-VL model controls
PDF, voice, and model selection in one place.

From page to playback

I built a self-hosted application that reconstructs PDF pages with local vision models, organizes the text, and narrates it with Kokoro. The browser interface brings document, voice, and model selection into one workflow.

The application supports a processing queue, chapter navigation, playback controls, and M4B audiobook downloads. It remains in active development, especially around reading order and complex page layouts.

Keep exploring.