We are excited to share that the latest Renku release (2.18.0 and 2.19.0) is out, and it extends Renku beyond interactive sessions. Renku’s data connectors and code integrations make it quick to start an exploratory analysis, but as projects mature and tasks run longer, the interactive session becomes a constraint. This release addresses that head-on, and much more!
Highlights
Run jobs. Create a Job launcher right next to your Session launchers. Run a project you’ve been working in as a job (same repo, same environment, same data, just add the command), or bring an external Docker image and run it on Renku compute. A single launcher can run many jobs at once. To learn more, check the docs.
Custom environments from private code repositories. Build images for sessions and jobs from private repos, not just public ones. No more ad-hoc public repositories just to build an image. (Note: only collaborators with read access to the repo can launch from it.) Big thanks to our collaborators at Idiap! 
Coding agents in your sessions. Bring Claude, Codex or Pi into a session by building the environment from code (add a project.toml). Renku also picks up your agent config folders (.claude,.pi,.codex). See an example project.
Dedicated pages for data connectors. Every data connector now has its own page: edit settings, explore metadata, and see which projects use it.
Read more in our blog post: Renku Release | July 2026 • Renku Jobs, Coding Agents & Custom Environments from Private Repos | Renku Blog
Watch the pre-release demo: https://youtu.be/JX_gU1LbxKY
As always, we’d love your feedback! Reply here or ping us on the forum.