Alexandria Audiobook Generator
Multi-voice AI audiobook generator
Description
A single machine voice reading a novel puts listeners to sleep, and human narration is expensive. Alexandria uses an LLM to annotate speakers and tone, then assigns each character a voice to produce a fully voiced, multi-voice audiobook.
Built on Qwen3-TTS, it supports voice cloning, voice design, LoRA training and per-line style control, exporting to MP3, chaptered M4B or an Audacity multi-track project.
Script annotation:LLM identifies speakers and tone.
Multi-voice:A voice per character.
Clone and design:Voice cloning, design and LoRA training.
Exports:MP3, chaptered M4B and Audacity multitrack.
Built on Qwen3-TTS, it supports voice cloning, voice design, LoRA training and per-line style control, exporting to MP3, chaptered M4B or an Audacity multi-track project.
Features
Script annotation:LLM identifies speakers and tone.
Multi-voice:A voice per character.
Clone and design:Voice cloning, design and LoRA training.
Exports:MP3, chaptered M4B and Audacity multitrack.
