General2 min read
Wiki.js 3.0 Beta
The beta for Wiki.js v3 is now available.
The beta for Wiki.js v3 is now available.
This is a complete rewrite of both the backend and frontend, using the latest technologies. Most of the major requests for v2 would have required significant changes to the code base (e.g. multi-sites, high-availability, live collaboration, etc.). There was also the issue of relying significantly on third-party libraries which introduced many breaking changes over time. This made it very difficult to continue with the v2 code base going forward. Wiki.js v3 is built to rely on the least possible number of dependencies, granting greater freedom for future updates.
On the frontend, v3 is now using Vue 3 and ditched the Vuetify framework in favor of a custom UI framework (inspired by the excellent Quasar framework). The app is now a full SPA (Single-Page Application) for seamless navigation between pages and other sections. This allows for a much improved editing experience as well. Page properties can now be changed without the need to enter the editor. Speaking of the editor, the markdown editor is now using the amazing Monaco engine, the same engine as Visual Studio Code. The live preview is an exact replica of the final render and saving pages is incredibly faster.
For the backend, the GraphQL API has been replaced with a OpenAPI-compliant REST API. While GraphQL was a good experiment and brings some advantages, it ended up being too confusing for most users and made integrations with other applications more difficult. The new REST API is fully documented and accessible via the /_api endpoint. The permission system is now clear and consistent. The HTTP server is now using Fastify while the authentication modules are custom rather than relying on often outdated Passport.js modules.
Regarding server requirements, Wiki.js v3 needs Node.js 26.x and PostgreSQL 16.x (or later). By focusing on a single database engine, development is greatly simplified and allows for more advanced features that are not possible with other engines. PostgreSQL is the far superior option and made advanced search, high-availability, live collaboration and more, possible directly via the database.
Blog posts focusing on specific new features are coming in the future.
This is a beta release. You shouldn’t install it in production or for critical scenarios. Bugs are to be expected and data loss is unlikely but possible.
Try it today by running the Docker container image.
Ask questions on Discord and report any bugs on GitHub Discussions.