Weaviate
Apply Gateco access policies to Weaviate hybrid BM25+vector search without modifying your Weaviate schema.
Search capabilities
| Vector (ANN) | Keyword (ranked) | Keyword (filter) | Hybrid (native) | Grep (substring) | Grep (regex) |
|---|---|---|---|---|---|
Weaviate cloud: <50ms p95 on standard tier. Policy overhead <25ms p95.
Weaviate is an AI-native vector database with GraphQL API, BM25 full-text search, and first-class hybrid search. Gateco connects to Weaviate as a Tier 2 retrieval connector, enforcing attribute-based and group-based policies on all supported search modes.
Weaviate's native hybrid search combines dense vector similarity with BM25 keyword ranking. Gateco passes the alpha value through to Weaviate's hybrid engine, so you get fully ranked hybrid results with Gateco's policy enforcement on top.
Weaviate object properties are available as inline metadata for policy evaluation — Gateco can read classification and ownership directly from the Weaviate object payload without maintaining a separate sidecar store.
Grep on Weaviate uses the Like operator with wildcard patterns — substring matching only. Regular expression grep is not supported on this connector.
Sample search config
{
"class_name": "Document"
}Sample policy
{
"name": "Internal content access",
"effect": "allow",
"rules": [{
"conditions": [
{"field": "resource.classification", "operator": "in", "value": ["internal", "public"]},
{"field": "principal.groups", "operator": "contains", "value": "employees"}
]
}],
"selectors": [{"connector_type": "weaviate"}]
}Policy conditions reference resource.* and principal.* fields. Policy reference →
Frequently asked questions
- Does Gateco support Weaviate multi-tenancy?
- Weaviate multi-tenancy (tenant-isolated classes) and Gateco policy enforcement are complementary. Gateco enforces policies within a class — you can combine Weaviate multi-tenancy for data isolation with Gateco policies for fine-grained attribute-based access control within each tenant.
- Can I use Weaviate generative search with Gateco?
- Gateco's grounded answers feature handles policy-aware answer synthesis (POST /api/answers/execute). Weaviate's native generative module runs server-side before Gateco applies policies. Use Gateco's answer synthesis for policy-enforced answer generation.
- Why does Weaviate grep not support regex?
- Weaviate's Like operator supports wildcard patterns (* and ?) but not regular expressions. Requests with search_mode=grep and pattern_type=regex are rejected with a clear error message for this connector.
Ready to connect Weaviate?
Follow the step-by-step setup guide or talk to the team for help with your specific configuration.