Skip to content

Roadmap

What's planned after v0.3.0.

These are plans, not promises. Priorities can change, and some items may land differently or not at all. The working plan is in SPEC.md (section 11, build phases), progress is in TODO.md, and what changed in each release is in the CHANGELOG.

Product work comes first: Area O1 should be easy to install and useful out of the box before the research work grows.

Coming next

  • export and purge commands. Export your case in one go, and delete a case cleanly. Until then, see Deleting a case.
  • More criteria profiles. Beyond O-1A and EB-1A, contributed by the community. You can already add your own as YAML in a workspace's profiles/ folder.
  • An adapter for Claude's dated-file export, alongside the current Claude and ChatGPT conversations.json imports.

Later in the plan

  • Session context packs and MCP memory tools. Richer ways for an agent to load what's relevant to a task from your claim memory.
  • An evaluation harness. Measured results for the claim memory (three systems, five scenarios, metrics and ablations), with a research-portfolio profile and fixture workspace. Area O1 makes no performance claims until this exists.
  • A Docker image and a GitHub Actions template for running jobs.
  • A short video walking through onboarding with a fictional persona.
  • Contributor guides for new connectors, profiles and opportunity scans.
  • Release artifacts: evaluation results and a Zenodo DOI with each release, and a documented PROV-JSON export.

Have an idea?

Open an issue or a discussion on GitHub. See Contributing if you'd like to build it.