Vellichor ghcr.io/woodscode/vellichor-web:latest https://github.com/woodscode/vellichor-web/pkgs/container/vellichor-web bridge bash false https://github.com/woodscode/vellichor-web/issues https://github.com/woodscode/vellichor-web Self-hosted web app that turns your written stories and ebooks into narrated audiobooks using the Kokoro-82M TTS model. Multi-voice character casting, background ambience, chaptered M4B + per-chapter MP3 output, and auto-export to an Audiobookshelf library. Optional NVIDIA GPU acceleration (requires the Nvidia Driver plugin); falls back to CPU. Optional AI "Smart cast" needs a separate Ollama container — set Ollama URL below; without it the rule-based Quick detect is used. MediaApp:Books Productivity: http://[IP]:[PORT:7777]/ https://raw.githubusercontent.com/woodscode/vellichor-web/main/unraid/icon.png --runtime=nvidia Turn stories & ebooks into narrated audiobooks with Kokoro-82M TTS. Multi-voice casting, ambience, M4B/MP3, Audiobookshelf export. GPU optional. 7777 /mnt/user/appdata/vellichor/data llama3.2:3b 99 100 all compute,utility