Roadmap
JJHub is constantly evolving. Here is a look at what we are currently building and what is planned for the future.Shipped & Feature Complete
The core JJHub platform is available in our closed alpha, providing:- Core jj Platform: Stable Change IDs, first-class conflicts, and seamless stacked changes.
- Landing Requests: Our AI-native replacement for pull requests, built for stacked changes.
- TypeScript Workflows: Unified AI and CI definitions (
.jjhub/workflows). - Programmable Gates: Control exactly how code is validated and merged.
- Agent Sandboxes: Secure runtime environments for executing AI operations natively.
Coming Soon
Web UI
A full-featured web interface for managing repositories, reading code, handling landing requests, and interacting with integrated agent workflows.Local Workflow Execution
A first-class capability to run your TypeScript-defined.jjhub/workflows entirely on your local machine. Test CI steps, run agent evaluations, and debug complex pipelines locally without needing to push to the cloud.
Facade Monorepos
Get the best of both worlds. Combine multiple distinct repositories (polyrepos) into a single “facade” monorepo. This provides the unified operational benefits of a monorepo for CI/CD and AI agents, while maintaining the clean separation of concerns and approachability of smaller, separate repositories for outside contributors.Advanced Programmable Issue & PR Workflows
Detailed, programmatic control over how issues and landing requests (PRs) can be created. Define specific rules and checks that must be proven before a submission is accepted.- Example: Reject an issue unless the user verifiably had an AI model confirm that a proper, reproducible minimum reproduction of the bug was provided.
- Example: Require specific security checks or sign-offs based on the files touched in a landing request.
First-class Smithers Support
Native integration with Smithers for managing autonomous, 24/7 running agents directly inside your JJHub repositories.OSS Support Payments
Empower open-source maintainers. Optionally gate feature requests, issues, or specific agent tasks behind direct payments to support the project and cover AI inference costs.Note: This roadmap is subject to change based on user feedback and technical discoveries.