Reference
get_ product_ intelligence_ batch
Up to 5 products in one call: results in input order, honest rate accounting.
Argument
1–5 product names. The bound is schema-enforced (minItems/maxItems), so well-behaved MCP clients reject an oversized batch before it reaches the server — which re-checks it anyway.
curl -s -X POST "https://services.buyary.com/v1/product-intelligence/batch" \
-H "Authorization: Bearer YOUR_BUYARY_KEY" \
-H "Content-Type: application/json" \
-d '{"product_names": ["iPhone 17 Pro Max", "totally unknown gadget"]}'Results
Each entry is exactly the single-call response shape, in input order. A name that doesn’t confidently resolve returns its own limited-coverage entry with did-you-mean candidates — one miss never fails the batch.
Rate accounting
A batch of N counts as N lookups against your key’s rate limit and usage metering — not one call. Batching saves round trips; it is not a quota loophole. A batch that would exceed your remaining rate budget fails with a tool error asking for fewer names or a short wait, before any lookups run.
detail=compact is not yet supported on the batch tool — batch entries always return the full shape.