system lane
PDF-to-Speech Audiobooks
A Python tool that converts PDF documents into audiobooks, using Google Gemini for text cleanup and speech synthesis, available as both a CLI and a Flask web app.
pipeline
- data (covered)
- model (not covered)
- eval (not covered)
- serve (covered)
- monitor (not covered)
PythonFlaskGemini APISQLAlchemy
linksI built and open-sourced this as a personal project. It extracts the text from a PDF and uses Gemini to clean up the OCR text, and then it generates the audio with Gemini’s text-to-speech. Long documents are processed in chunks so that they stay within the API quota limits.
The web app has an upload and progress interface, an in-browser audio player, and job history, with persistent storage through SQLAlchemy against SQLite or PostgreSQL.