KosmoKrator

database

Pinecone MCP Integration for LangChain

Connect Pinecone to LangChain through the local KosmoKrator MCP gateway with scoped tools, credentials, and write policy.

7 functions 5 read 2 write Bearer token auth
Submit Pinecone feedback View package source

Connect Pinecone to LangChain

Bridge LangChain agents to local KosmoKrator integration tools through MCP or headless CLI calls.

Use the MCP gateway when the agent should discover tools, or wrap kosmo integrations:call for fixed chains. The gateway is local, scoped to this integration, and starts with --write=deny so LangChain can inspect read-capable tools without receiving write access by default.

Pinecone MCP Config for LangChain

Keep the gateway scoped to the integration and operation class needed by the chain.

{
  "mcpServers": {
    "kosmokrator-pinecone": {
      "type": "stdio",
      "command": "kosmo",
      "args": [
        "mcp:serve",
        "--integration=pinecone",
        "--write=deny"
      ]
    }
  }
}

Run the Gateway Manually

kosmokrator mcp:serve --integration=pinecone --write=deny

Why Use KosmoKrator Here

Scoped tools

Expose only Pinecone instead of a broad multi-service tool list.

Local credentials

Reuse credentials already configured for the KosmoKrator CLI and Lua runtime.

Write policy

Start read-only, then opt into ask or allow for trusted workspaces.

Pinecone Tools Visible to LangChain

LangChain sees stable MCP tool names generated from the Pinecone integration catalog.

MCP toolSource functionTypeDescription
integration__pinecone__pinecone_list_indexes pinecone.pinecone_list_indexes Read List all vector indexes in your Pinecone project. Returns index names, dimensions, metrics, and status.
integration__pinecone__pinecone_get_index pinecone.pinecone_get_index Read Get detailed information about a specific Pinecone vector index, including its dimension, metric, host URL, and status.
integration__pinecone__pinecone_create_index pinecone.pinecone_create_index Write Create a new serverless vector index in Pinecone. Specify the index name, vector dimension, and similarity metric (cosine, euclidean, or dotproduct).
integration__pinecone__pinecone_upsert_vectors pinecone.pinecone_upsert_vectors Write Upsert vectors into a Pinecone index using an index host URL.
integration__pinecone__pinecone_query_vectors pinecone.pinecone_query_vectors Read Search for similar vectors in a Pinecone index using a query embedding.
integration__pinecone__pinecone_list_collections pinecone.pinecone_list_collections Read List all collections in your Pinecone project. Collections are static snapshots of indexes used for backups or creating new indexes.
integration__pinecone__pinecone_get_current_user pinecone.pinecone_get_current_user Read Get information about the currently authenticated Pinecone user, including email, name, and project details.

Related Pinecone Pages

Pinecone CLIUse direct JSON commands instead of MCP discovery. Pinecone MCPFull gateway reference and MCP tool names. KosmoKrator for LangChainClient-level gateway setup.