Roadmap

April 26, 2026 · View on GitHub

DeepEye is currently focused on stabilizing the public development preview.

Current Focus

Improve setup, documentation, and contributor onboarding.
Strengthen sandbox, generated-code execution, and runtime-control boundaries.
Consolidate the workflow model around session -> turn -> draft -> run -> artifact.
Keep backend/core/frontend checks reliable in CI.
Continue reducing legacy paths and duplicated state handling.
Make artifact rendering depend on a stable typed artifact protocol.
Add repeatable local quality gates and security scans.

Document API groups and auth/session concepts.
Provide a small reproducible demo dataset and workflow walkthrough.
Improve production hardening guidance for deployments beyond local development.
Add more focused tests around workflow node execution and artifact persistence.
Improve dependency and container image update automation.
Introduce the artifact protocol normalization layer described in docs/rfcs/artifact_protocol.md.
Start behavior-preserving splits of the largest backend node implementation files.

Define release/versioning policy.
Publish architecture diagrams and threat model notes.
Split reusable workflow node implementations if they outgrow the backend runtime layer.
Add benchmark or evaluation examples for common analytical workflows.