Sandai's reference context — project or loop

When you ask Sandai to draft or sharpen something, it doesn't work from the open form alone — it pulls in relevant artefacts you've already created to ground its suggestions. A User Persona draws on your Stakeholders, a Lean Canvas draws on your Problem and Next Big Idea, and so on.

Reference context decides how wide Sandai casts that net.

The two settings

• Whole project (default) — Sandai may reference artefacts from every loop. Best when your loops are small and share one theme, so learning from an earlier loop should inform the current one.
• Current loop only — Sandai references only the loop being worked in. Best when each loop is a distinct, self-contained pass — a separate bet or a different audience — and you don't want one loop's detail bleeding into another.

This only affects what Sandai reads for grounding. It never changes who can see or edit your work, and it doesn't change the artefacts themselves.

Setting it

Open Project settings → Details. The Sandai reference context dropdown is set by the project owner. The choice is per project, so you can keep one project project-wide and another loop-scoped at the same time.

Note: whole-project analyses — like Sandai's review across all your completed loops — always read everything by design. This setting governs the drafting context for tools, not that cross-loop analysis.