Package model and spec context to hand to AI.
Open a specification and use Copy specification context for AI to package it into one self-contained bundle — prose plus resolved model attachments — ready to paste into AI Architect, Cursor, or your agent. Exporting context makes the input explicit, scoped, and repeatable, rather than relying on screenshots or a hand-written summary.
What the bundle includes
The export resolves the spec’s live embeds into machine-readable form:
| Included | Form |
|---|---|
| Goal, scope, and implementation notes | Spec prose |
| Objects diagram | Live model slice from YAML |
| Workflow sequence | Sequence diagram for the relevant process |
| ERD | Data-model slice |
| Contracts | gRPC / OpenAPI embeds |
Because the bundle is built from the model, it always reflects the current state — not a stale snapshot pasted earlier.
When to export
- Working in an external AI client without live access.
- Sharing a precise, frozen context for reproducibility.
- Scoping AI to a specific subtree, not the whole project.
Flow
Choose the scope
Pick the object, subtree, or spec that matters for the task.
Export the context
Produce a snapshot (model JSON, expanded MDX, or a rendered diagram).
Provide it to AI
Attach or paste it as the grounding for your prompt.
Apply results back
Bring accepted changes into the model through normal review.
Scoped snapshots
A focused neighborhood export usually beats a whole-project dump: it gives AI exactly the relevant structure without noise.
Pair an exported diagram with the structured model JSON. The diagram conveys shape; the JSON gives AI precise, machine-readable structure.
Last updated on