qdrant

# Qdrant Vector Database Base URL: `http://localhost:6333` (default). Override with `QDRANT_URL` env var. All examples use curl. Replace `localhost:6333` with the actual Qdrant endpoint. ## Quick Reference | Task | Endpoint | |------|----------| | List collections | `GET /collections` | | Create collection | `PUT /collections/{name}` | | Delete collection | `DELETE /collections/{name}` | | Collection info | `GET /collections/{name}` | | Upsert points | `PUT /collections/{name}/points` | | Search | `POST /collections/{name}/points/search` | | Scroll (paginate) | `POST /collections/{name}/points/scroll` | | Get point | `GET /collections/{name}/points/{id}` | | Delete points | `POST /collections/{name}/points/delete` | | Count points | `POST /collections/{name}/points/count` | | Health check | `GET /healthz` | ## Collections ### List all collections ```bash curl -s http://localhost:6333/collections | jq ``` ### Create collection ```bash curl -s -X PUT http://localhost:6333/collections/my_collection \ -H "Content-Type: application/json" \ -d '{ "vectors": { "size": 384, "distance": "Cosine" } }' | jq ``` Distance metrics: `Cosine`, `Euclid`, `Dot`, `Manhattan`. ### Create collection with named vectors ```bash curl -s -X PUT http://localhost:6333/collections/my_collection \ -H "Content-Type: application/json" \ -d '{ "vectors": { "content": {"size": 384, "distance": "Cosine"}, "title": {"size": 128, "distance": "Cosine"} } }' | jq ``` ### Collection info ```bash curl -s http://localhost:6333/collections/my_collection | jq ``` Key fields: `vectors_count`, `points_count`, `segments_count`, `status` (green/yellow/red). ### Delete collection ```bash curl -s -X DELETE http://localhost:6333/collections/my_collection | jq ``` ### Check collection exists ```bash curl -sI http://localhost:6333/collections/my_collection -o /dev/null -w "%{http_code}" # 200 = exists, 404 = not found ``` ## Points ### Upsert points (insert or update) ```bash curl -s -X PUT http://localhost:6333/collections/my_collection/points \ -H "Content-Type: application/json" \ -d '{ "points": [ { "id": 1, "vector": [0.1, 0.2, 0.3, ...], "payload": {"title": "Example", "category": "test", "score": 0.95} }, { "id": 2, "vector": [0.4, 0.5, 0.6, ...], "payload": {"title": "Another", "category": "prod", "score": 0.87} } ] }' | jq ``` Point IDs can be integers or UUIDs. ### Upsert with named vectors ```bash curl -s -X PUT http://localhost:6333/collections/my_collection/points \ -H "Content-Type: application/json" \ -d '{ "points": [ { "id": 1, "vector": { "content": [0.1, 0.2, ...], "title": [0.3, 0.4, ...] }, "payload": {"title": "Example"} } ] }' | jq ``` ### Get point by ID ```bash curl -s http://localhost:6333/collections/my_collection/points/1 | jq ``` ### Get multiple points ```bash curl -s -X POST http://localhost:6333/collections/my_collection/points \ -H "Content-Type: application/json" \ -d '{"ids": [1, 2, 3]}' | jq ``` ### Delete points by ID ```bash curl -s -X POST http://localhost:6333/collections/my_collection/points/delete \ -H "Content-Type: application/json" \ -d '{"points": [1, 2, 3]}' | jq ``` ### Delete points by filter ```bash curl -s -X POST http://localhost:6333/collections/my_collection/points/delete \ -H "Content-Type: application/json" \ -d '{ "filter": { "must": [ {"key": "category", "match": {"value": "obsolete"}} ] } }' | jq ``` ### Count points ```bash # Total count curl -s -X POST http://localhost:6333/collections/my_collection/points/count \ -H "Content-Type: application/json" \ -d '{}' | jq '.result.count' # Count with filter curl -s -X POST http://localhost:6333/collections/my_collection/points/count \ -H "Content-Type: application/json" \ -d '{ "filter": { "must": [{"key": "category", "match": {"value": "test"}}] }, "exact": true }' | jq '.result.count' ``` ## Search ### Basic vector search ```bash curl -s -X POST http://localhost:6333/collections/my_collection/points/search \ -H "Content-Type: application/json" \ -d '{ "vector": [0.1, 0.2, 0.3, ...], "limit": 10, "with_payload": true, "with_vector": false }' | jq ``` ### Search with filter ```bash curl -s -X POST http://localhost:6333/collections/my_collection/points/search \ -H "Content-Type: application/json" \ -d '{ "vector": [0.1, 0.2, 0.3, ...], "limit": 5, "with_payload": true, "filter": { "must": [ {"key": "category", "match": {"value": "prod"}} ], "must_not": [ {"key": "archived", "match": {"value": true}} ] } }' | jq ``` ### Search with score threshold ```bash curl -s -X POST http://localhost:6333/collections/my_collection/points/search \ -H "Content-Type: application/json" \ -d '{ "vector": [0.1, 0.2, 0.3, ...], "limit": 10, "score_threshold": 0.7, "with_payload": true }' | jq ``` ### Search with named vector ```bash curl -s -X POST http://localhost:6333/collections/my_collection/points/search \ -H "Content-Type: application/json" \ -d '{ "vector": { "name": "content", "vector": [0.1, 0.2, 0.3, ...] }, "limit": 10, "with_payload": true }' | jq ``` ### Scroll (paginate through all points) ```bash # First page curl -s -X POST http://localhost:6333/collections/my_collection/points/scroll \ -H "Content-Type: application/json" \ -d '{ "limit": 100, "with_payload": true, "with_vector": false }' | jq # Next page (use next_page_offset from previous response) curl -s -X POST http://localhost:6333/collections/my_collection/points/scroll \ -H "Content-Type: application/json" \ -d '{ "limit": 100, "offset": NEXT_PAGE_OFFSET, "with_payload": true }' | jq ``` ### Recommend (find similar to existing points) ```bash curl -s -X POST http://localhost:6333/collections/my_collection/points/recommend \ -H "Content-Type: application/json" \ -d '{ "positive": [1, 2], "negative": [3], "limit": 10, "with_payload": true }' | jq ``` ## Filtering Filters use `must` (AND), `should` (OR), `must_not` (NOT) clauses. ### Condition types ```json // Exact match {"key": "city", "match": {"value": "Berlin"}} // Match any {"key": "color", "match": {"any": ["red", "blue", "green"]}} // Match except {"key": "status", "match": {"except": ["deleted", "archived"]}} // Range (numeric) {"key": "price", "range": {"gte": 10.0, "lte": 100.0}} // Range (dates as strings) {"key": "created_at", "range": {"gte": "2025-01-01T00:00:00Z"}} // Geo bounding box {"key": "location", "geo_bounding_box": { "top_left": {"lat": 52.52, "lon": 13.38}, "bottom_right": {"lat": 52.50, "lon": 13.42} }} // Geo radius {"key": "location", "geo_radius": { "center": {"lat": 52.52, "lon": 13.405}, "radius": 1000.0 }} // Has key (field exists) {"has_id": [1, 2, 3]} // Is empty (field missing or null) {"is_empty": {"key": "optional_field"}} // Is null {"is_null": {"key": "nullable_field"}} // Nested filter {"nested": {"key": "tags", "filter": { "must": [{"key": "tags[].name", "match": {"value": "important"}}] }}} // Values count {"values_count": {"key": "tags", "gte": 2}} ``` ### Combined filter example