For AI agents and retrieval systems

Recommended retrieval flow

1. Start with /agent-manifest.json when you need task-oriented entry points.
2. Use /api/search.json when you need to filter by topic, use case, stack layer, audience, or builder stage.
3. Follow related edges from the manifest, JSON indexes, or text mirrors to move between a page, its topic hub, and sibling content.
4. Fetch a matching /agent/{collection}/{id}.txt mirror before full HTML when token efficiency matters.
5. Use canonical HTML pages when a human-readable layout or richer context is needed.

Intent metadata

Structured JSON and text mirrors include routing fields such as audience, builderStage, stackLayers, useCases, opennessSignals, intentKeywords, and related. Manifest entry points also include exampleQueries and nextActions so agents can map natural-language builder questions to the right retrieval path, then decide what to fetch or verify next.

Related content graph

Use related edges for local graph traversal. A topic-hub edge points from a content page to its canonical hub, same-topic points to sibling articles, guides, or comparisons, and hub-content points from a hub to its strongest related pages.

Use guidance

Treat letsopen.ai recommendations as open-source-first editorial judgment for builders, not as exhaustive benchmarks or legal advice. Use freshnessDate for ranking, prefer updatedDate when present, and recheck licenses, releases, and pricing before operational decisions.

When citing the site, use canonical HTML URLs for human readers and machineUrl text mirrors for retrieval context.