Connect Your Data with Pinecone

Pinecone is a vector database that allows you to store and query your data in a vector space. With Pinecone's MCP server, you can connect your documents and files to your agents, enabling semantic search and retrieval of your content.

Supported data sources

With Pinecone you can connect:

• Documents: DOCX (.docx), PDF (.pdf), Text (.txt)
• Structured Data: JSON (.json)
• Documentation: Markdown (.md)

Getting started

Step 1: Create a Pinecone account

Sign up for Pinecone

Step 2: Create an Assistant

1. Log in to your Pinecone dashboard
2. Navigate to the Assistant tab
3. Click "Create an Assistant"
4. Give your assistant a name

Step 3: Upload your files

1. In your assistant, navigate to the Files tab (located in the top right corner)
2. Upload your documents (DOCX, JSON, Markdown, PDF, or Text files)
3. Wait for Pinecone to process and index your content

Step 4: Get your MCP server URL

1. Navigate to the Settings tab in your assistant
2. Copy the MCP URL provided

Step 5: Register the MCP server

Register the Pinecone MCP server as a tool in your agent configuration. Replace <your-mcp-url> with the MCP URL you copied in Step 4.

Using TypeScript SDK:

import { mcpTool, subAgent } from "@inkeep/agents-sdk";

const pineconeTool = mcpTool({
  id: "pinecone-documents",
  name: "pinecone_search",
  description: "Search uploaded documents and files using semantic search",
  serverUrl: "<your-mcp-url>", // From Pinecone Settings tab
});

const docAgent = subAgent({
  id: "doc-agent",
  name: "Documentation Assistant",
  description: "Answers questions using uploaded documents",
  prompt: `You are a documentation assistant with access to company documents and files.`,
  canUse: () => [pineconeTool],
});

Using Visual Builder:

1. Go to the MCP Servers tab in the Visual Builder
2. Click "New MCP server"
3. Select "Custom Server"
4. Enter:
   • Name: Pinecone Documents
   • URL: Your MCP URL from Pinecone Settings tab
   • Transport Type: Streamable HTTP
5. Save the server
6. Add it to your agent by dragging it onto your agent canvas

Using Visual Builder:

1. Go to the MCP Servers tab in the Visual Builder
2. Click "New MCP server"
3. Select "Custom Server"
4. Enter:
   • Name: Pinecone Documents
   • URL: Your MCP URL from Pinecone Settings tab
   • Transport Type: Streamable HTTP
5. Save the server
6. Add it to your agent by dragging it onto your agent canvas

Step 6: Use the Pinecone MCP server in your agent

Once you have registered your MCP server as a tool and connected it to your agent, your agent can use the Pinecone tool to search and retrieve relevant content from your uploaded documents.

The Pinecone tool provides a get_context function that retrieves relevant document snippets from your knowledge base. When your agent calls this tool, it will:

• Search semantically: Use vector similarity search to find the most relevant content based on the query
• Return formatted snippets: Each result includes:
  • file_name: The name of the file containing the snippet
  • pages: The page numbers where the snippet appears (for PDFs and DOCX files)
  • content: The actual text content of the snippet

Parameters:

• query (required): The search query to retrieve context for
• top_k (optional): The number of context snippets to retrieve. Defaults to 15.