Roadmap¶

This is directional and priorities and scope will change.

More storage remotes¶

• Google Cloud Storage, alongside the existing Azure Blob and S3 remotes.

CI integration¶

• A GitHub Action that wraps the push and promote flow.

Running evals¶

• pytest-xdist support, to run evals in parallel across workers.

Dashboard¶

• The pooling window as a project setting and as a control in the dashboard.

• Runner details on every attempt, and link templates that turn a recorded trace id into a link to an observability platform.

• Images in inputs and outputs, stored once by content hash and rendered inline.