Matchora

Matchora scans a video library or ingests a list of titles, searches metadata APIs, and writes an NFO catalog. Candidates are ranked with SequenceMatcher. The browser UI is an admin console.

Shipped providers: TVMaze and Jikan (no key). OMDb and TMDB are optional and need API keys.

Install

Download matchora-<version>-linux-amd64.tar.gz from GitHub Releases. Unpack it. The archive root is matchora/ (binary, config/, public/, LICENSE).

Run

From that directory:

./matchora

The process listens on http.addr from config/default.yaml (shipped as port 7680). Open the admin console on that host and port.

Runtime data lives in data/ next to the binary: session job files, secrets, an optional config.yaml overlay, and the NFO catalog. Pass -config to load a different default.yaml.

Library and keys

The folder picker stays inside browse_root, which defaults to data/. Point it at a real library by setting browse_root in data/config.yaml (same keys as default.yaml).

TVMaze and Jikan need no key. Set OMDb and TMDB keys in the admin secrets panel, or in data/secrets.

Use

• Scan a path under the browse root. Matchora groups files into titles, then searches providers.
• Ingest a CSV or JSON list of titles (optional year, type, season, episode, IMDb id).
• High-confidence hits auto-match. Close scores stay manual until you pick a candidate.
• Matched titles are written under data/catalog as NFO trees with posters.

From source

• Developers build and package the same tarball with ./build/run (rebuild and package). See docs/dev/workflow.md.
• Design notes: architecture, match, gui.