Search in Collections
Query vector collections with natural language questions or technical keywords. Returns ranked passages with source document references, page numbers, and similarity scores.
POST/v1/collections/{collection_id}/search
Search content in a collection
Executes hybrid retrieval over the collection shards and returns the most relevant chunks.
| Parameter | Type | Description |
|---|---|---|
| queryrequired string | string | The natural language search question or keyword text. |
| top_k integer | integer | Number of matching chunks to return (default 5, max 50). |
| score_threshold float | float | Minimum similarity threshold between 0.0 and 1.0. |
cURL Requestbash
curl https://api.plannix.web.id/v1/collections/col_883a9/search \
-H "Authorization: Bearer ${PLANNIX_API_KEY}" \
-H "Content-Type: application/json" \
-d '{
"query": "indemnification clauses",
"top_k": 5
}'Hybrid ranking and threshold parameters
Pass hybrid_alpha: 0.7 to weight dense semantic vector similarity vs 0.3 lexical BM25 matching.
Metadata filtering expressions
Filter chunks by document attributes using MongoDB-style operators ($eq, $in, $gt, $and).