kavita-lightnovel-metadata-fetcher/README.md

kavita-lightnovel-metadata-fetcher

Pulls metadata (summary, tags, genres, characters, staff, score, cover, links, related series) for light novels from MangaBaka, enriched with MyAnimeList and AniList data, and writes it back to a Kavita server through its REST API.

No file mover, no ComicInfo.xml — the source of truth is Kavita itself. Series are discovered via the Kavita library API.

Features

• Match every series in one or more Kavita libraries against MangaBaka and persist the match in matches.json (editable via the web UI).
• Update metadata for a single series or all matched series at once. Updates are diff-based:
  • Locked fields in Kavita are never overwritten.
  • List fields (tags, genres, characters, writers, …) are merged: new items are added, removed items are dropped.
  • Cover images are only re-uploaded when MangaBaka's cover URL actually changed.
• Characters and authors are synced to Kavita Person records (image, description, MAL/AniList id) via Kavita's /api/Person endpoints.
• MangaBaka relationships (sequel / prequel / spin-off / …) are mirrored as Kavita series relationships, and every related series that exists in Kavita is added to a shared collection.

Environment

Variable	Default	Description
KAVITA_URL	—	Base URL of the Kavita server, e.g. http://kavita:5000
KAVITA_API_KEY	—	API key from Kavita user settings
LIBRARY_IDS	(empty)	Default libraries (CSV of ids). Empty = pick in WebUI.
LANGUAGE	en	Series language ISO code (used for language field)
REQUEST_TIMEOUT	30	HTTP timeout in seconds
MATCH_PATH	/config/matches.json	Where to persist the match cache
WEB_HOST	0.0.0.0	Bind host for the Flask UI
WEB_PORT	8080	Bind port for the Flask UI

Running locally

pip install -r requirements.txt
KAVITA_URL=http://localhost:5000 KAVITA_API_KEY=... python main.py

Then open http://localhost:8080/.

Docker

docker compose -f docker-compose.prod.yml up -d