Pinecone Endpoint

Pinecone endpoints connect Eve to Pinecone indexes for governed vector search, retrieval, and AI application workflows.

Endpoint Kind

Use pinecone as the endpoint kind when creating this endpoint through the API or dashboard.

What Eve Uses Pinecone Endpoints For

Run governed vector similarity search.
Expose retrieval workflows to Adam and approved tools.
Centralize Pinecone API key handling.
Combine vector retrieval with LLM, SQL, document, and service endpoints.

Basic Endpoint Shape

json

{
  "endpoint": "pinecone-prod",
  "kind": "pinecone",
  "config": {
    "read_conn": {
      "api_key": "secret",
      "environment": "us-east-1-aws",
      "index": "customer-docs"
    }
  },
  "description": "Customer document retrieval index"
}

Create it with the endpoint API:

bash

curl -sS -X POST "$EDEN/endpoints" \
  -H "$AUTH_HEADER" \
  -H "Content-Type: application/json" \
  -d '{
    "endpoint": "pinecone-prod",
    "kind": "pinecone",
    "config": {
      "read_conn": {
        "api_key": "secret",
        "environment": "us-east-1-aws",
        "index": "customer-docs"
      }
    },
    "description": "Customer document retrieval index"
  }'

Configuration Fields

Field	Purpose
`api_key`	Pinecone API key.
`environment`	Pinecone environment or region.
`index`	Index name.

Operating Notes

Treat retrieved documents as sensitive application data.
Scope API keys to the indexes and operations Eve needs.
Monitor latency, result counts, and prompt injection risk when retrieval results feed LLM workflows.