Agent-readable docs index: /llms.txt. Full docs in one file: /llms-full.txt. Download /docs.zip to grep all markdown files locally.

Dashboard

The Dashboard is your home view. When your workspace is connected, the stat tiles aggregate across every review in the workspace — total findings, blockers, needs-review, and findings still awaiting disposition — so the numbers tell workspace truth, not just the review you have open.

Needs your attention

Findings the engine is not confident enough to assert, plus everything still open in the loaded review, are collected here. Jump straight to the Review queue to disposition them — accept, reject, or resolve each finding with an owner and comment.

Recent reviews

Your latest reviews across all review types — drawings, Material Certs, and P&ID — with finding and open counts for each. Use All documents to jump to the document library. When the workspace summary is unavailable (or in demo mode), the dashboard scopes itself to the review you have loaded and says so.

Review confidence

Each review carries a confidence indicator. High-confidence findings are asserted directly; low-confidence findings route to the review queue so a person makes the call.