Search Tuning Now Applies to Every API Call and Integration, and embed_and_search Matches Your Hosted Search Page Exactly
Search tuning used to stop at the edge of the hosted search page. It now travels with every search made through the API and through the client integrations, so results returned in code are the same results you see on your own search page.
01 One tuned pipeline everywhere
Everything you configure in Index Settings → Search Tuning — the semantic↔lexical balance, field weights, minimum match, search mode, vector candidate pool and content quality boost — now applies automatically to every search and every RAG answer made through the integrations. It is the same tuned pipeline as your hosted search page, not an approximation of it.
02 embed_and_search returns what your search page returns
The one-shot hybrid search endpoint now runs the platform's full tuned pipeline. Your index's saved Search Tuning applies automatically, and natural-language questions are processed exactly as they are on your search page: stopwords are cleaned for keyword matching, while the full phrase goes to the semantic embedder. Results returned via the API are identical to your search page.
03 Per-call overrides
Each knob can be overridden per call from code through the new tuning parameter, and every tuning knob accepts per-request overrides on the endpoint.
| Parameter | Controls |
|---|---|
fw_* | Field weights |
mm | Minimum match |
search_mode | Search mode |
vector_topk | Vector candidate pool |
quality_boost | Content quality boost |
04 Package versions
- LangChain and LlamaIndex. Updated to carry saved tuning and the
tuningparameter. - Haystack. v0.2.5.
- Laravel Scout. v0.2.5.
- MCP. v0.2.6.