Best Free & Open Source Alternatives to Eagle

Monbooru is a single-binary, self-hosted gallery for a personal image collection, built around the booru conventions that anime and AI-art collectors already use: tag categories with aliases and implications, a general-to-explicit rating scale with a configurable SFW ceiling, ratings, favourites, saved searches, collections and a search syntax with wildcards, OR and exclusions. Images, video, animated GIFs and CBZ/ZIP archives stay as ordinary files and folders on disk, while a per-gallery SQLite database holds tags, relations and thumbnails. Removing monbooru leaves your files exactly where they were.

It is aimed squarely at local media libraries rather than phone-camera rollups. Stable Diffusion metadata from A1111/Forge and ComfyUI (prompts, models, seeds, full workflows) is read and shown per image, and local ONNX taggers such as WD14 SwinV2, animetimm EVA02, JoyTag, Camie v2 and PixAI can label the collection on CPU or GPU without sending anything off the machine. A file watcher picks up new, moved and deleted files within seconds, SHA-256 hashing keeps one file from becoming two entries, and declared relations plus near-duplicate detection with a side-by-side review session handle alternates, version chains and duplicates.

The core app never touches the internet. Downloading from boorus and reverse image lookup (IQDB, SauceNAO, Hydrus Public Tag Repository) happen through the optional monloader companion, with a browser extension (monsender) for sending the page you are on. Compared with cloud photo services such as Google Photos or iCloud Photos, there is no storage subscription and no upload, but also no mobile apps or automatic camera backup. Compared with a paid desktop catalogue like Eagle, you get booru-style tagging, an inbox review workflow and a documented REST API, and you host it yourself instead of licensing a seat.

Highlights

• Tag-based gallery with folder tree, ratings, favourites, saved searches and collections
• Local ONNX auto-tagging (WD14, EVA02, JoyTag, Camie v2, PixAI) on CPU or GPU, with per-model tag provenance
• Stable Diffusion and ComfyUI metadata, including full generation workflows
• Manga reader for CBZ/ZIP archives, plus video, GIF and animated hover previews on the grid
• Batch operations across a selection or a whole search: delete, move, tag, auto-tag, set collection
• Multiple galleries per instance, each with its own filesystem and database, with export/import
• REST API under /api/v1 with scoped tokens, OpenAPI spec and built-in docs; plugin and theme support
• Optional password login; intended for local network use, not public exposure

Tropy is a free, open-source desktop application built for people who photograph archival material and then have to live with the results: thousands of images still named IMG_5261.JPG, with the box, folder and collection they came from living only in a notebook. It was created at the Roy Rosenzweig Center for History and New Media and is now developed jointly with the Luxembourg Centre for Contemporary and Digital History and Digital Scholar, the nonprofit behind Zotero and Omeka.

The workflow is built around research sources rather than camera rollups. Photos are imported by reference, combined into items (the three pages of a letter become one item), described with custom metadata templates, tagged, and grouped into lists. Templates can mix in external vocabularies, including the Dublin Core, RDF and Europeana vocabularies that ship with the app plus any .ttl or .n3 vocabulary you import, so fields map onto archival description practice instead of a fixed schema. Item view adds notes and transcriptions in a rich text editor, selections cut out a detail of a page, and the built-in photo tools only go as far as rotate, crop and zoom to make a document legible. Search runs across metadata, tags and notes.

A project is a single .tpy file backed by SQLite, the storage format the Library of Congress recommends for preservation, and your photos are linked rather than modified, so removing Tropy leaves the originals untouched. Metadata and notes export to JSON-LD, CSV, a ZIP archive or directly to Omeka S, with a CSL export for Zotero; IIIF manifests can be imported, and plugins handle custom formats. Compared with paid photo catalogues such as Adobe Lightroom or Eagle, Tropy adds archival templates, transcriptions and selections but deliberately skips ratings, develop modules and cloud sync, and it is explicitly not a citation manager or a writing environment. Collaboration across machines is still manual: put the project on an external or cloud drive and open it read-only when the research is finished.

Highlights

• Items, lists, tags and full-text search over metadata, notes and tags
• Custom metadata templates with Dublin Core, RDF and Europeana vocabularies, plus imported .ttl/.n3 vocabularies
• Notes and transcriptions, selections within an image, and basic rotate/crop/zoom
• Export to JSON-LD, CSV, ZIP archive, Omeka S and CSL for Zotero; IIIF manifest import
• Plugin system for custom imports and exports, with a SQLite-backed project file
• Native builds for macOS, Windows and Linux, including Flatpak, Homebrew, winget and AUR