# Pinecone MCP for AI Agents AI Agent Connect

> Pinecone MCP lets your AI agent talk directly to your vector database. Query embeddings, manage collections, and pull performance stats through natural conversation. It turns your vector store into an accessible knowledge base for your agent without you having to write custom scripts.

## Overview
- **Category:** loved-by-devs
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_cMLbuhs0rnfwH86edlO4MVqnOLr94ybdUvJavFsC/ai-agent-connect
- **Tags:** semantic-search, vector-embeddings, knowledge-graph, high-performance, ai-infrastructure

## Description

This Connector gives your AI client direct access to your Pinecone environment. Instead of jumping between your terminal and the Pinecone console to check on your data, you can just ask your agent to do it for you. You can have your agent look up specific vector details, check how full your pods are, or clean up old data on the fly. It makes managing a complex vector space feel like a conversation rather than a series of API calls. Whether you're building a RAG system or just need to audit your storage, this connection puts the power of your vector store right into your chat interface. You can find this and thousands of other tools in the Vinkius catalog to build out your entire AI stack.

## Tools

### delete_vectors
Removes specific vectors from an index to free up space or clear data. This is useful for maintaining storage hygiene.

### describe_index
Shows the configuration details and topology of a specific index. Use this to verify your pod settings and dimensions.

### fetch_vectors
Retrieves specific vectors using their unique IDs for precise lookups. It's perfect for pinpointing exact records.

### get_index_stats
Pulls real-time health checks and capacity limits for your pods. This helps you monitor your storage usage at a glance.

### list_collections
Lists all index collections to help you organize snapshot arrays. Use it to see your grouped data structures.

### query_vectors
Finds and returns the most similar vectors and metadata for a given input. This is the core tool for semantic search.

### list_indexes
Shows every index currently existing in your Pinecone environment. It gives you a clear view of your entire vector store.

## Prompt Examples

**Prompt:** 
```
Check the vector count stats for the index named `document-embeddings`.
```

**Response:** 
```
### Index Stats: `document-embeddings` 

| Metric | Value |
| :--- | :--- |
| **Vector Count** | 45,920 |
| **Pod Capacity** | 90% Full |
| **Dimensions** | 1536 |

The index is approaching capacity. You might want to consider scaling your pods soon.
```

**Prompt:** 
```
Delete all vectors belonging to the user ID 'auth-abc123' namespace.
```

**Response:** 
```
✅ **Deletion Successful**

I've wiped all vectors associated with the `auth-abc123` namespace from your index. The space has been cleared.
```

**Prompt:** 
```
List all existing collections created in my Pinecone environment.
```

**Response:** 
```
You have the following active collections:

*   `backup-q1-2026` (Snapshot)
*   `knowledge-base-staging` (Snapshot)

Let me know if you need more details on any of these.
```

## Capabilities

### Find similar vectors
Search your database to find the most relevant embeddings based on your input.

### Fetch specific data
Pull precise vector records using their unique IDs.

### View index list
See every index currently existing in your environment at once.

### Check index settings
View the configuration details and topology parameters of a specific index.

### Monitor usage stats
Pull real-time health checks and capacity limits for your pods.

### List collections
See all snapshot arrays and grouped data in your space.

### Remove vectors
Delete specific data points from an index to free up space.

## Use Cases

### Testing RAG retrieval
An engineer asks the agent to find similar vectors for a query to see if the retrieved context is relevant for the prompt.

### Configuration audit
An architect asks the agent to describe the index configuration to verify the mathematical dimensions and pod settings.

### Targeted retrieval
A developer uses fetch_vectors to pull specific data points by their unique IDs for a precise data check.

### Capacity planning
A data custodian asks for the vector count stats to see if they need to scale their production environment.

### Data cleanup
A developer tells the agent to remove all records belonging to a specific user ID to comply with a deletion request.

## Benefits

- Check RAG relevance quickly. Use query_vectors to see if your agent is pulling the right context.
- Monitor storage without the console. Use get_index_stats to see your pod capacity in real time.
- Track your snapshots. Use list_collections to see all your grouped data arrays.
- See your whole environment. Use list_indexes to get a bird's eye view of your vector store.
- Faster debugging. You can check your topology parameters without hunting through menus.
- Easier cleanup. You can wipe specific user data or test records with a single command.

## How It Works

The bottom line is you get a conversational interface for your vector database.

1. Subscribe to the Pinecone MCP in the Vinkius marketplace.
2. Add your Pinecone API key to your AI client configuration.
3. Ask your agent to run queries or pull stats on your vector stores.

## Frequently Asked Questions

**Can I use the Pinecone MCP to manage my vector database?**
Yes, this Connector connects your AI client directly to Pinecone. You can query data, check stats, and manage your indexes using natural language.

**How do I check my vector capacity with Pinecone?**
You can simply ask your agent to pull the usage stats. It will check your pod capacity and vector counts in real time.

**Can my AI agent delete vectors for me?**
Yes, you can give your agent the command to delete specific vectors. This is great for clearing out test data or handling user deletion requests.

**Is the Pinecone MCP good for testing RAG?**
It's perfect for RAG. You can ask your agent to run queries and see what context it retrieves, helping you debug relevance without writing scripts.

**How do I see all my indexes in Pinecone?**
Just ask your agent to list your indexes. It will return a list of all the indexes currently in your environment.

**Can I use this to check my index configurations?**
Yes, your agent can describe any specific index to show you its configuration, topology, and other parameters instantly.

**Can the AI execute raw vector similarity searches?**
Yes, absolutely. Once you supply the raw semantic embedding coordinates (normally a float array generated previously), the LLM can funnel it through the `query_vectors` tool. The Pinecone DB will process this and return the top-K closest vector matches along with embedded metadata.

**How do I check my remaining vector storage capacity?**
It's extremely simple. Just ask the connected AI agent to 'Get the index stats'. It will internally call `get_index_stats` against the specified index namespace, returning total vector count and physical dimensionality limits to your chat window.

**Is it safe to delete vectors dynamically using the chat terminal?**
Yes, but with standard precautions. The `delete_vectors` tool operates exactly as the official SDK. As long as you maintain clear contextual scopes and ID filtering in your prompts, the execution is purely deterministic and secure.