How to integrate Pinecone MCP with CrewAI

GET STARTED FOR FREE
Framework Integration Gradient
Pinecone Logo
CrewAI Logo
divider

Introduction

This guide walks you through connecting Pinecone to CrewAI using the Composio tool router. By the end, you'll have a working Pinecone agent that can query all vectors similar to user question, upsert document embeddings into a namespace, delete vectors from the archive index through natural language commands.

This guide will help you understand how to give your CrewAI agent real control over a Pinecone account through Composio's Pinecone MCP server.

Before we dive in, let's take a quick look at the key ideas and tools involved.

TL;DR

Here's what you'll learn:
  • Get a Composio API key and configure your Pinecone connection
  • Set up CrewAI with an MCP enabled agent
  • Create a Tool Router session or standalone MCP server for Pinecone
  • Build a conversational loop where your agent can execute Pinecone operations

What is CrewAI?

CrewAI is a powerful framework for building multi-agent AI systems. It provides primitives for defining agents with specific roles, creating tasks, and orchestrating workflows through crews.

Key features include:

  • Agent Roles: Define specialized agents with specific goals and backstories
  • Task Management: Create tasks with clear descriptions and expected outputs
  • Crew Orchestration: Combine agents and tasks into collaborative workflows
  • MCP Integration: Connect to external tools through Model Context Protocol

What is the Pinecone MCP server, and what's possible with it?

The Pinecone MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Pinecone account. It provides structured and secure access so your agent can perform Pinecone operations on your behalf.

Supported Tools & Triggers

Tools
Cancel Bulk ImportTool to cancel a bulk import operation in Pinecone.
Configure IndexTool to configure an existing Pinecone index, including pod type, replicas, deletion protection, and tags.
Create BackupTool to create a backup of a Pinecone index for disaster recovery and version control.
Create IndexTool to create a Pinecone index with specified configuration.
Create Index with Embedding ModelTool to create a Pinecone index with integrated embedding model for automatic vectorization.
Create Index from BackupTool to create an index from a backup.
Create NamespaceTool to create a namespace within a serverless Pinecone index.
Delete IndexTool to permanently delete a Pinecone index.
Delete NamespaceTool to permanently delete a namespace from a serverless index.
Describe BackupTool to retrieve detailed information about a specific backup.
Describe Bulk ImportTool to describe a specific bulk import operation in Pinecone.
Describe Index StatsTool to get index statistics including vector count per namespace, dimensions, and fullness.
Describe Restore JobTool to get detailed information about a specific restore job in Pinecone.
Generate EmbeddingsTool to generate vector embeddings for input text using Pinecone's hosted embedding models.
Get Model InformationTool to retrieve detailed information about a specific model hosted by Pinecone.
List Bulk ImportsTool to list all recent and ongoing bulk import operations in Pinecone.
List CollectionsTool to list all collections in a Pinecone project (pod-based indexes only).
List Index BackupsTool to list all backups for a specific Pinecone index.
List IndexesTool to list all indexes in a Pinecone project.
List Available ModelsTool to list all available embedding and reranking models hosted by Pinecone.
List NamespacesTool to list all namespaces in a serverless Pinecone index.
List Project BackupsTool to list all backups for indexes in a Pinecone project.
List Restore JobsTool to list all restore jobs for a project with pagination support.
List VectorsTool to list vector IDs in a Pinecone serverless index.
Query VectorsTool to perform semantic search within a Pinecone index using a query vector.
Rerank DocumentsTool to rerank documents by semantic relevance to a query.
Search Records in NamespaceTool to search records within a Pinecone namespace using text, vector, or ID query.
Start Bulk ImportTool to start an asynchronous bulk import of vectors from object storage (S3, GCS, or Azure Blob Storage) into a Pinecone index.
Update VectorTool to update a vector in Pinecone by ID.
Upsert Records to NamespaceTool to upsert text records into a Pinecone namespace.

What is the Composio tool router, and how does it fit here?

What is Tool Router?

Composio's Tool Router helps agents find the right tools for a task at runtime. You can plug in multiple toolkits (like Gmail, HubSpot, and GitHub), and the agent will identify the relevant app and action to complete multi-step workflows. This can reduce token usage and improve the reliability of tool calls. Read more here: Getting started with Tool Router

The tool router generates a secure MCP URL that your agents can access to perform actions.

How the Tool Router works

The Tool Router follows a three-phase workflow:

  1. Discovery: Searches for tools matching your task and returns relevant toolkits with their details.
  2. Authentication: Checks for active connections. If missing, creates an auth config and returns a connection URL via Auth Link.
  3. Execution: Executes the action using the authenticated connection.

Step-by-step Guide

Prerequisites

Before starting, make sure you have:
  • Python 3.9 or higher
  • A Composio account and API key
  • A Pinecone connection authorized in Composio
  • An OpenAI API key for the CrewAI LLM
  • Basic familiarity with Python

Getting API Keys for OpenAI and Composio

OpenAI API Key
  • Go to the OpenAI dashboard and create an API key. You'll need credits to use the models, or you can connect to another model provider.
  • Keep the API key safe.
Composio API Key
  • Log in to the Composio dashboard.
  • Navigate to your API settings and generate a new API key.
  • Store this key securely as you'll need it for authentication.

Install dependencies

bash
pip install composio crewai crewai-tools[mcp] python-dotenv
What's happening:
  • composio connects your agent to Pinecone via MCP
  • crewai provides Agent, Task, Crew, and LLM primitives
  • crewai-tools[mcp] includes MCP helpers
  • python-dotenv loads environment variables from .env

Set up environment variables

bash
COMPOSIO_API_KEY=your_composio_api_key_here
USER_ID=your_user_id_here
OPENAI_API_KEY=your_openai_api_key_here

Create a .env file in your project root.

What's happening:

  • COMPOSIO_API_KEY authenticates with Composio
  • USER_ID scopes the session to your account
  • OPENAI_API_KEY lets CrewAI use your chosen OpenAI model

Import dependencies

python
import os
from composio import Composio
from crewai import Agent, Task, Crew
from crewai_tools import MCPServerAdapter
import dotenv

dotenv.load_dotenv()

COMPOSIO_API_KEY = os.getenv("COMPOSIO_API_KEY")
COMPOSIO_USER_ID = os.getenv("COMPOSIO_USER_ID")

if not COMPOSIO_API_KEY:
    raise ValueError("COMPOSIO_API_KEY is not set")
if not COMPOSIO_USER_ID:
    raise ValueError("COMPOSIO_USER_ID is not set")
What's happening:
  • CrewAI classes define agents and tasks, and run the workflow
  • MCPServerHTTP connects the agent to an MCP endpoint
  • Composio will give you a short lived Pinecone MCP URL

Create a Composio Tool Router session for Pinecone

python
composio_client = Composio(api_key=COMPOSIO_API_KEY)
session = composio_client.create(user_id=COMPOSIO_USER_ID, toolkits=["pinecone"])

url = session.mcp.url
What's happening:
  • You create a Pinecone only session through Composio
  • Composio returns an MCP HTTP URL that exposes Pinecone tools

Initialize the MCP Server

python
server_params = {
    "url": url,
    "transport": "streamable-http",
    "headers": {"x-api-key": COMPOSIO_API_KEY},
}

with MCPServerAdapter(server_params) as tools:
    agent = Agent(
        role="Search Assistant",
        goal="Help users search the internet effectively",
        backstory="You are a helpful assistant with access to search tools.",
        tools=tools,
        verbose=False,
        max_iter=10,
    )
What's Happening:
  • Server Configuration: The code sets up connection parameters including the MCP server URL, streamable HTTP transport, and Composio API key authentication.
  • MCP Adapter Bridge: MCPServerAdapter acts as a context manager that converts Composio MCP tools into a CrewAI-compatible format.
  • Agent Setup: Creates a CrewAI Agent with a defined role (Search Assistant), goal (help with internet searches), and access to the MCP tools.
  • Configuration Options: The agent includes settings like verbose=False for clean output and max_iter=10 to prevent infinite loops.
  • Dynamic Tool Usage: Once created, the agent automatically accesses all Composio Search tools and decides when to use them based on user queries.

Create a CLI Chatloop and define the Crew

python
print("Chat started! Type 'exit' or 'quit' to end.\n")

conversation_context = ""

while True:
    user_input = input("You: ").strip()

    if user_input.lower() in ["exit", "quit", "bye"]:
        print("\nGoodbye!")
        break

    if not user_input:
        continue

    conversation_context += f"\nUser: {user_input}\n"
    print("\nAgent is thinking...\n")

    task = Task(
        description=(
            f"Conversation history:\n{conversation_context}\n\n"
            f"Current request: {user_input}"
        ),
        expected_output="A helpful response addressing the user's request",
        agent=agent,
    )

    crew = Crew(agents=[agent], tasks=[task], verbose=False)
    result = crew.kickoff()
    response = str(result)

    conversation_context += f"Agent: {response}\n"
    print(f"Agent: {response}\n")
What's Happening:
  • Interactive CLI Setup: The code creates an infinite loop that continuously prompts for user input and maintains the entire conversation history in a string variable.
  • Input Validation: Empty inputs are ignored to prevent processing blank messages and keep the conversation clean.
  • Context Building: Each user message is appended to the conversation context, which preserves the full dialogue history for better agent responses.
  • Dynamic Task Creation: For every user input, a new Task is created that includes both the full conversation history and the current request as context.
  • Crew Execution: A Crew is instantiated with the agent and task, then kicked off to process the request and generate a response.
  • Response Management: The agent's response is converted to a string, added to the conversation context, and displayed to the user, maintaining conversational continuity.

Complete Code

Here's the complete code to get you started with Pinecone and CrewAI:

python
from crewai import Agent, Task, Crew, LLM
from crewai_tools import MCPServerAdapter
from composio import Composio
from dotenv import load_dotenv
import os

load_dotenv()

GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY")
COMPOSIO_API_KEY = os.getenv("COMPOSIO_API_KEY")
COMPOSIO_USER_ID = os.getenv("COMPOSIO_USER_ID")

if not GOOGLE_API_KEY:
    raise ValueError("GOOGLE_API_KEY is not set in the environment.")
if not COMPOSIO_API_KEY:
    raise ValueError("COMPOSIO_API_KEY is not set in the environment.")
if not COMPOSIO_USER_ID:
    raise ValueError("COMPOSIO_USER_ID is not set in the environment.")

# Initialize Composio and create a session
composio = Composio(api_key=COMPOSIO_API_KEY)
session = composio.create(
    user_id=COMPOSIO_USER_ID,
    toolkits=["pinecone"],
)
url = session.mcp.url

# Configure LLM
llm = LLM(
    model="gpt-5",
    api_key=os.getenv("OPENAI_API_KEY"),
)

server_params = {
    "url": url,
    "transport": "streamable-http",
    "headers": {"x-api-key": COMPOSIO_API_KEY},
}

with MCPServerAdapter(server_params) as tools:
    agent = Agent(
        role="Search Assistant",
        goal="Help users with internet searches",
        backstory="You are an expert assistant with access to Composio Search tools.",
        tools=tools,
        llm=llm,
        verbose=False,
        max_iter=10,
    )

    print("Chat started! Type 'exit' or 'quit' to end.\n")

    conversation_context = ""

    while True:
        user_input = input("You: ").strip()

        if user_input.lower() in ["exit", "quit", "bye"]:
            print("\nGoodbye!")
            break

        if not user_input:
            continue

        conversation_context += f"\nUser: {user_input}\n"
        print("\nAgent is thinking...\n")

        task = Task(
            description=(
                f"Conversation history:\n{conversation_context}\n\n"
                f"Current request: {user_input}"
            ),
            expected_output="A helpful response addressing the user's request",
            agent=agent,
        )

        crew = Crew(agents=[agent], tasks=[task], verbose=False)
        result = crew.kickoff()
        response = str(result)

        conversation_context += f"Agent: {response}\n"
        print(f"Agent: {response}\n")

Conclusion

You now have a CrewAI agent connected to Pinecone through Composio's Tool Router. The agent can perform Pinecone operations through natural language commands.

Next steps:

  • Add role-specific instructions to customize agent behavior
  • Plug in more toolkits for multi-app workflows
  • Chain tasks for complex multi-step operations

How to build Pinecone MCP Agent with another framework

OpenAI Agents SDK

OpenAI Agents SDK

Use Pinecone MCP with OpenAI Agents SDK

Claude Agent SDK

Claude Agent SDK

Use Pinecone MCP with Claude Agent SDK

Google ADK

Google ADK

Use Pinecone MCP with Google ADK

LangChain

LangChain

Use Pinecone MCP with LangChain

Vercel AI SDK

Vercel AI SDK

Use Pinecone MCP with Vercel AI SDK

Mastra AI

Mastra AI

Use Pinecone MCP with Mastra AI

LlamaIndex

LlamaIndex

Use Pinecone MCP with LlamaIndex

Claude Code

Claude Code

Use Pinecone MCP with Claude Code

Codex

Codex

Use Pinecone MCP with Codex

FAQ

What are the differences in Tool Router MCP and Pinecone MCP?

With a standalone Pinecone MCP server, the agents and LLMs can only access a fixed set of Pinecone tools tied to that server. However, with the Composio Tool Router, agents can dynamically load tools from Pinecone and many other apps based on the task at hand, all through a single MCP endpoint.

Can I use Tool Router MCP with CrewAI?

Yes, you can. CrewAI fully supports MCP integration. You get structured tool calling, message history handling, and model orchestration while Tool Router takes care of discovering and serving the right Pinecone tools.

Can I manage the permissions and scopes for Pinecone while using Tool Router?

Yes, absolutely. You can configure which Pinecone scopes and actions are allowed when connecting your account to Composio. You can also bring your own OAuth credentials or API configuration so you keep full control over what the agent can do.

How safe is my data with Composio Tool Router?

All sensitive data such as tokens, keys, and configuration is fully encrypted at rest and in transit. Composio is SOC 2 Type 2 compliant and follows strict security practices so your Pinecone data and credentials are handled as safely as possible.

Used by agents from

Context
Letta
glean
HubSpot
Agent.ai
Altera
DataStax
Entelligence
Rolai
Context
Letta
glean
HubSpot
Agent.ai
Altera
DataStax
Entelligence
Rolai
Context
Letta
glean
HubSpot
Agent.ai
Altera
DataStax
Entelligence
Rolai

Never worry about agent reliability

We handle tool reliability, observability, and security so you never have to second-guess an agent action.

Book a call↗Get started for free