Pinecone
Add per-principal access control to Pinecone that static metadata filters cannot provide.
Search capabilities
| Vector (ANN) | Keyword (ranked) | Keyword (filter) | Hybrid (native) | Grep (substring) | Grep (regex) |
|---|---|---|---|---|---|
Pinecone ANN: <100ms p95 at 1M vectors (standard pod). Gateco policy overhead <25ms p95.
Pinecone is a managed vector database optimized for high-throughput approximate nearest-neighbor (ANN) search. Gateco connects to Pinecone as a Tier 1 connector, supporting both ingestion and retrieval with policy enforcement.
Pinecone's native metadata filtering lets you pre-filter by document properties stored at index time. Gateco adds dynamic principal context: the same Pinecone index can return different results to different users based on their IDP-synced groups, department, and attributes — without per-query metadata re-indexing.
Retroactive registration lists all vector IDs from Pinecone (via Pinecone's list API) and registers them as Gateco resources. You can apply policies to your existing index without re-embedding.
Pinecone supports vector-only search in Gateco v1. Keyword, hybrid, and grep modes are not available for this connector. For ranked keyword or hybrid search alongside Pinecone, consider routing those query types to a separate pgvector or OpenSearch connector.
Sample search config
{
"index_name": "knowledge-base",
"namespace": "production"
}Sample policy
{
"name": "Employee knowledge access",
"effect": "allow",
"rules": [{
"conditions": [
{"field": "resource.classification", "operator": "in", "value": ["internal", "public"]},
{"field": "principal.groups", "operator": "contains", "value": "employees"}
]
}],
"selectors": [{"connector_type": "pinecone"}]
}Policy conditions reference resource.* and principal.* fields. Policy reference →
Frequently asked questions
- How is Gateco different from Pinecone built-in metadata filters?
- Pinecone metadata filters operate on static key-value pairs stored at index time. Gateco adds dynamic principal context: it evaluates policies against IDP-synced principal attributes (groups, department, custom attributes) that change as users join or leave teams. Gateco also provides an audit trail, policy versioning, and a cross-connector policy model — one policy can cover Pinecone and your other connectors simultaneously.
- Does Pinecone support retroactive registration?
- Yes. Gateco uses Pinecone's list API to enumerate vector IDs in the index, then registers each as a gated resource. This works on existing indexes without re-embedding. Metadata for each resource is stored in Gateco's sidecar store (inline resolution from Pinecone payload is also supported).
- Can I use keyword or hybrid search with Pinecone?
- Not in Gateco v1. Pinecone's architecture does not support BM25 keyword ranking or grep patterns. For keyword or hybrid retrieval alongside Pinecone, route those queries to a separate pgvector or OpenSearch connector.
Ready to connect Pinecone?
Follow the step-by-step setup guide or talk to the team for help with your specific configuration.