CatalogueCanvas is a self-hosted catalogue server. You upload ZIP files containing your work, organize them into collections, add metadata and descriptions, and share curated portfolios as shareable pages.
Unlike a repository or organized folder structure, CatalogueCanvas treats your work as FAIR data: Findable through search and tagging, Accessible with metadata and preview generation, Interoperable across formats and libraries, and Reusable through portfolios and structured metadata.
This organization contains the source code and documentation for CatalogueCanvas:
- CatalogueCanvas — the application
- cataloguecanvas-website — documentation and showcase site
- cataloguecanvas-homeassistant-addon — Home Assistant add-on
- Ingest — upload ZIP files as items; images convert to WebP previews automatically
- Organize — group items into collections, mark favorites, search
- Annotate — edit titles, tags, notes; use a local vision model to generate descriptions
- Publish — create portfolios and share them at
/p/<slug>links - Store — back up the database and all files; use multiple storage libraries
Digital artists cataloguing a body of work and publishing portfolios. Generative coders keeping rendered images and source code in the same item. Designers organizing asset sets. Studios and self-hosters running a catalogue on their own hardware.
Docker is the easiest way to run it, even on a low-powered machine. Bare-metal installs also work on Linux, macOS, and Windows, but you'll need Python 3.11+, Node.js 22+, and uv.
See the full installation guide for details.
GNU Affero General Public License v3.0