BVector
Configuration
BVector can be configured via environment variables or a JSON config file mounted into the container.
Environment variables
| Variable | Required | Description |
|---|---|---|
| BVECTOR_AUTH_TOKEN | required | Authentication token for API access. Obtain from the BangDB dashboard. |
| BVECTOR_LOG_LEVEL | optional | Log verbosity: debug | info | warn | error. Default: info |
| BVECTOR_MAX_CONNECTIONS | optional | Maximum concurrent client connections. Default: 100 |
| BVECTOR_DATA_DIR | optional | Path to the persistent data directory. Default: /data |
| BVECTOR_PORT | optional | HTTP API port. Default: 18080 |
| BVECTOR_RESOURCE_PORT | optional | Resource service port. Default: 18082 |
Config file
Mount a JSON config file at /etc/bvector/config.json as an alternative to environment variables:
config.json
{
"server": {
"port": 18080,
"resourcePort": 18082,
"maxConnections": 100,
"logLevel": "info"
},
"storage": {
"dataDir": "/data",
"distanceMetric": "cosine"
},
"index": {
"hnswM": 16,
"hnswEfConstruction": 200,
"hnswEfSearch": 50
}
}HNSW index tuning
The HNSW parameters control the speed/accuracy tradeoff. Adjust these based on your dataset size and query latency requirements.
hnswMNumber of bidirectional links per node. Higher = better recall, more memory. Range: 4–64. Default: 16.
hnswEfConstructionDynamic candidate list size during index build. Higher = better quality, slower ingestion. Default: 200.
hnswEfSearchDynamic candidate list size at query time. Higher = better recall, slower queries. Default: 50.