A library health tool for Plex, Kodi, Emby and Jellyfin users: scan your offline movie library, find broken or incomplete metadata and artwork, and fix them.

Download for Windows: get MediaClinic-0.18.4-win64.zip (about 19 MB) from the v0.18.4 release. Unzip it, open the MediaClinic folder and double-click MediaClinic.exe. No Python needed. The SHA-256 checksum is in the release notes.

v0.19.0 (multiple libraries, ID verification, Auto Repair, image-status fixes) is published as source only: run it from Python (see Quick start). The next Windows build comes with v0.19.1.

Windows SmartScreen: the file is not code-signed, so Windows may show "Windows protected your PC". Click More info, then Run anyway. You can also run the app from source (see Quick start) and read every line of the code first.

MediaClinic expects one subfolder per movie, the layout used by Plex, Kodi, Emby and Jellyfin. It checks Kodi .nfo and Emby/Jellyfin movie.xml files, so it also suits Plex libraries that use an NFO add-on. It checks every artwork file, metadata file and video, and shows one colour-coded health table, red for errors and yellow for warnings. From the same table you can repair the common problems in place.

Every check is a health rule you can switch on or off, or tune, in Settings → Health Rules.

• Normalize Genres and Normalize Sources: written to both NFO and XML
• Sync Ratings from TMDb and OMDb (you supply your own free API keys)
• Verify IMDb/TMDb IDs against TMDb (title in any translation, year ±1) with a sortable report
• Auto Repair for one movie, a selection or the whole library, with a dry-run preview: IDs, ratings, genres, year, titles and XML language; artwork is never touched
• Fetch Poster, Fanart and Backdrops from TMDB with thumbnails, or extract backdrop frames from the video with FFmpeg
• Copy poster.jpg → folder.jpg where the two differ, for one movie, a selection or the whole library
• Export Movie List to CSV, text or Excel, with column presets (including Identify: IMDB and TMDB links plus outline) and the option to export only the filtered view
• Several libraries in one table (Libraries… dialog and library filter); an offline drive keeps its cached rows
• Large libraries: parallel scanning, clickable Errors/Warnings counters, Problems only, a column chooser, Library Statistics and a Duplicate Movies Report
• Safe edits: every NFO/XML is backed up silently and written atomically, with automatic backup folder maintenance; rows refresh automatically after each change
• Keyboard shortcuts, jump-to-letter, sortable columns, a Quick Scan that re-reads only what changed, remembered window sizes and positions, rotating logs

From the Windows download, just run MediaClinic.exe. From source:

pip install -r requirements.txt     # optional: pillow, openpyxl
python mediaclinic.py

Run it from the folder that contains the settings_*.py files. FFmpeg and FFprobe are optional but recommended (point Settings → Tools to their bin folder). Settings, your API keys and the last scan are stored in %APPDATA%\MediaMetadataClinic\, outside this repository; logs/ and backup/ are created next to the program. See docs/help for the in-app help pages.

This project is the clearest example of working with an AI under a written contract:

• docs/CLAUDE_RULES.md defines module boundaries, the save chain, naming rules and a safety checklist that every AI-edited change must respect
• settings_schema.json plus tests/settings_test_harness.py check the settings subsystem after every change
• tests/selftest_v0_19_0.py builds a synthetic library (NFO, XML, generated JPEGs and a real 3-second MKV) and runs about 285 end-to-end checks, from health rules to Auto Repair and the export dialogs (TMDb calls are simulated); it runs in CI under a virtual display and never touches your real settings
• The code is split into model, controller, context, dialog, export, help and genre-data modules so the AI can edit one part without breaking another
• About 17,500 lines of Python in eight modules; 32 earlier versions in archive/versions/, starting from a 12 KB folder scanner
• Changes per version are in CHANGELOG.md

Built with Claude (Anthropic) as the coding partner. I wrote the requirements and the revision prompts, tested every build on real data, and decided what to fix next. The archive/versions/ folder keeps every earlier release so the iteration history is visible.

Security note: the app stores any API keys you enter in a local settings file outside this repository. .gitignore excludes config and settings files so keys are never committed.

See CONTRIBUTING.md and SECURITY.md. Bug reports and ideas are welcome through the issue templates.

MIT. See LICENSE. This product uses the TMDB API but is not endorsed or certified by TMDB. OMDb data is subject to the OMDb API terms. FFmpeg is a separate download under its own license.

Luiz Junqueira - junqueira.ch - LinkedIn