mcp-pinecone mcp

v1.0.0 · MCP Tool · database · registry.pascalai.org

Pinecone vector database via REST API. Control plane: list_indexes, describe_index (requires indexName). Data plane: query, upsert, fetch, delete, list_vectors, describe_stats (requires indexHost). Required always: apiKey.

pineconevectorembeddingssimilaritydatabasecloud

Input Parameters

Parameter	Type	Description
operationrequired	string	Operation to perform. One of: `list_indexes`, `describe_index`, `query`, `upsert`, `fetch`, `delete`, `list_vectors`, `describe_stats`.
apiKeyrequired	string	Pinecone API key.
indexNameoptional	string	Index name for control-plane operations.
indexHostoptional	string	Index host URL for data-plane operations (e.g. https://my-index.svc.pinecone.io).
vectoroptional	string	Float array JSON for query.
topKoptional	integer	Number of top results to return (default: 10). Default: `10`.
namespaceoptional	string	Namespace for data operations.
filteroptional	string	JSON metadata filter.
idsoptional	string	JSON array of vector IDs for fetch/delete.
vectorsoptional	string	JSON array of {id,values,metadata?} for upsert.
includeValuesoptional	boolean	Include vector values in query results (default: false). Default: `False`.
includeMetadataoptional	boolean	Include metadata in query results (default: true). Default: `True`.
deleteAlloptional	boolean	Delete all vectors in namespace. Default: `False`.
limitoptional	integer	Max vectors for list_vectors (default: 100). Default: `100`.
paginationTokenoptional	string	Pagination token for list_vectors.

Output Fields

Field	Type	Description
indexes	array[object]
index	object
matches	array[object]
vectors	object
stats	object
ok	boolean
error	string

Examples

Query Pinecone index

// Input
{
  "operation": "query",
  "apiKey": "pc-xxx",
  "indexHost": "https://my-index.svc.pinecone.io",
  "vector": "[0.1,0.2,0.3]",
  "topK": 5
}

// Output
{
  "matches": [
    {
      "id": "v1",
      "score": 0.98
    }
  ]
}

Install & Discovery

Install

ppm install mcp-pinecone

Get JSON Schema

GET /v1/packages/mcp-pinecone/1.0.0/schema

Discover by keyword

GET /v1/mcp/discover?q=pinecone

Discovery hint: Install with ppm install mcp-pinecone or invoke remotely via POST /v1/invoke/mcp-pinecone on the MCP Service.

PascalAI Usage

uses toolslib;
var Tool := LoadTool('mcp-pinecone');
var R := Tool.Call(JsonObj(['operation','list_indexes','apiKey','pc-xxx']));
Writeln(R.ToJSON);