Agents

How do I resolve a fuzzy watch name to a canonical reference?

Call /api/references/resolve with the user's fuzzy text, then cite the highest-confidence canonical URL before fetching /api/references/{slug} for market context.

Resolver endpoint
/api/references/resolve
Canonical references
151
Best first query
q=Rolex Pepsi

Why resolve references before answering?

Fuzzy names like Pepsi, Jumbo or 5711 can map to multiple references. Resolution reduces citation drift.

What should an AI agent cite?

Cite the canonical URL returned in matches[].url, then fetch /api/references/{slug} for full context.