5bc69ad9ce
All checks were successful
Release / Test backend (push) Successful in 6m5s
Release / Test UI (push) Successful in 1m44s
Release / Build and push images (push) Successful in 7m6s
Release / Deploy to prod (push) Successful in 3m47s
Release / Deploy to homelab (push) Successful in 8s
Release / Gitea Release (push) Successful in 44s
Llama 4 Scout returns the model output directly as a JSON array in the
'response' field rather than as a plain string or [{"generated_text":"..."}].
The old fallback parsed it as [{"generated_text":""}] (Go JSON ignores
unknown fields) and silently returned an empty string — causing chapter-names
jobs to finish with results:null in the payload despite items_done matching
items_total.
Fix: only accept the generated_text fallback when the value is non-empty;
otherwise return the raw JSON bytes as a string so callers (parseChapterTitlesJSON
etc.) can extract what they need.
LibNovel
Self-hosted audiobook platform. Go backend + SvelteKit UI + MinIO/PocketBase/Meilisearch.
Requirements
- Docker + Docker Compose
- just
- Doppler CLI
Setup
doppler login
doppler setup # project=libnovel, config=prd
Usage
just up # start everything
just down # stop
just logs # tail all logs
just log backend # tail one service
just build # rebuild images
just restart # down + up
just secrets # view/edit secrets
Secrets
Managed via Doppler (project=libnovel, config=prd). No .env files.
To add or update a secret:
doppler secrets set MY_SECRET=value