# Couchbase (Vector & NoSQL) MCP for AI Agents AI Agent Connect

> Couchbase (Vector & NoSQL) MCP lets your AI agent talk directly to your Couchbase cluster. It handles N1QL queries, KNN vector searches, and full-text searches while letting you audit schemas and manage documents across buckets, scopes, and collections. It's the bridge between your unstructured data and your AI's reasoning.

## Overview
- **Category:** databases
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_wWaJnEVjNwe96vxf5IajtX1h1rZH9VkCF8KWnsvT/ai-agent-connect
- **Tags:** nosql, vector-search, knn, n1ql, embeddings, high-performance-computing

## Description

Managing a hybrid database environment usually means jumping between different tools to check your NoSQL structures or run complex queries. If you're building a RAG application, you might find yourself manually verifying how your vector indices are performing or digging through JSON payloads to see why a specific record isn't showing up. This Connector changes that by putting your entire Couchbase cluster inside your AI's reach. Instead of writing out long N1QL strings or manually checking bucket configurations, you can just ask your agent to find specific data patterns or explain your current schema. It handles the heavy lifting of navigating scopes and collections so you can stay focused on the logic of your app. Whether you're auditing existing indexes or prototyping new search features, it gives you a direct line to your data without the friction of constant context switching. It's one of the many ways Vinkius helps you connect your specialized infrastructure to the models you use every day. You get a clear view of your data organization and the ability to execute complex queries using plain language, making it a lot easier to debug your production environment or build out new features quickly.

## Tools

### execute_n1ql_query
Run N1QL queries to get specific JSON payloads from your buckets. It lets you perform complex data filtering using SQL-like logic.

### list_buckets
See all the routing spaces in your Couchbase DB. This helps you understand the high-level organization of your data.

### list_scopes
Find the limits and objects for your specific scopes and collections. It helps you see exactly what data is accessible.

### list_indexes
See all the active search indexes and rules attached to your data. This is vital for verifying your search configurations.

### vector_search
Map structural KNN vector similarities using your existing vector indexes. This is the core tool for similarity-based retrieval.

### get_document
Pull internal properties for specific Couchbase KV documents using their keys. It gives you an exact look at a single record.

### fts_search
Match query strings against your content trees using full-text search indexes. It handles structural text-based extraction.

## Prompt Examples

**Prompt:** 
```
Show me all the buckets in my Couchbase cluster.
```

**Response:** 
```
I've checked your cluster and found the following buckets:

| Bucket Name | Description |
| :--- | :--- |
| `travel-sample` | Sample travel data |
| `user-profiles` | Active user accounts |
| `product-catalog` | Inventory items

Which one would you like to explore further?
```

**Prompt:** 
```
Find products similar to these coordinates: [0.12, -0.5, 0.88]
```

**Response:** 
```
I've performed a KNN vector search on your `product-catalog` index. Here are the top 3 matches:

* **Wireless Pro Headphones** (Score: 0.98)
* **Noise Cancelling Earbuds** (Score: 0.92)
* **Bluetooth Headset** (Score: 0.89)

Would you like to see the full JSON metadata for the top result?
```

**Prompt:** 
```
Run a query to get all travel items under $100.
```

**Response:** 
```
I executed that N1QL query for you. Here are the results from the `travel-sample` bucket:

| Item | Price |
| :--- | :--- |
| Economy Flight | $85 |
| Local Tour | $45 |
| Train Pass | $90 |

I have the full JSON objects available if you need to see more details.
```

## Capabilities

### Run N1QL queries
Execute complex SQL-like queries against your JSON data using natural language.

### Perform KNN vector searches
Find similar items based on vector embeddings for RAG applications.

### Audit database schemas
See your buckets, scopes, and collections to understand your data layout.

### Fetch specific documents
Grab exact JSON records using unique keys for quick inspection.

### Execute full-text searches
Match query strings against your content trees using FTS indexes.

### Inspect search indexes
View all your registered indexes and verify vector definitions.

## Use Cases

### Debugging a RAG pipeline
An AI engineer notices poor retrieval results and asks the agent to use vector_search to check the similarity scores of the top 5 results.

### Auditing a new production environment
A data architect needs to verify the setup and asks the agent to use list_buckets and list_scopes to confirm the schema matches the staging plan.

### Quick JSON data extraction
A product manager wants to see the structure of a specific user profile and asks the agent to use get_document to pull the JSON.

### Verifying N1QL query performance
A DBA runs a complex execute_n1ql_query through the agent to see how many records are returned and check for any unexpected nulls.

## Benefits

- Faster RAG debugging because you can use vector_search to see exactly how your embeddings are being matched in real-time.
- Eliminate manual schema checks by using list_buckets and list_scopes to see your entire data organization instantly.
- Run complex data retrievals with execute_n1ql_query without having to write and test long SQL strings manually.
- Audit your search infrastructure quickly with list_indexes to verify your vector and FTS definitions are correct.
- Get instant access to specific data points using get_document to see the exact JSON structure of any record.
- Simplify content discovery by using fts_search to match text patterns across your entire database.

## How It Works

The bottom line is you get a direct conversation line to your NoSQL and vector data without touching a terminal.

1. Subscribe to the Connector on Vinkius.
2. Enter your Couchbase URL, username, and password.
3. Start asking your AI client to query data or search vectors.

## Frequently Asked Questions

**Can I use the Couchbase (Vector & NoSQL) MCP to manage my RAG application?**
Yes. It allows your AI agent to perform KNN vector searches and N1QL queries, making it much easier to debug and refine how your AI retrieves information from your database.

**Does the Couchbase (Vector & NoSQL) MCP work with my self-hosted cluster?**
It works with both Couchbase Capella and self-hosted clusters. You just need to provide your URL and credentials to give your agent access.

**Can I use this to see my database structure?**
Absolutely. You can ask the agent to list your buckets, scopes, and collections to get a clear picture of how your data is organized without running manual commands.

**Is the Couchbase (Vector & NoSQL) MCP good for auditing indexes?**
Yes, it's a great way to quickly enumerate all your search indexes and verify your vector definitions to ensure your search logic is set up correctly.

**Can my AI agent actually run N1QL queries?**
Yes, the Connector allows your agent to execute N1QL queries and return the JSON results directly in your chat, saving you from writing and testing queries manually.

**How do I connect my Couchbase data to my AI client using this?**
Once you subscribe on Vinkius, you just enter your database credentials. From there, your AI client can start querying your NoSQL and vector data immediately.

**Can my agent perform K-Nearest Neighbor (KNN) vector searches in Couchbase?**
Yes. Provide the search index name, the vector embedding array, and the number of results (k). The agent uses Couchbase's native vector capabilities to locate the most semantically similar documents in your cluster.

**How do I execute a N1QL query through the agent?**
Use the 'execute_n1ql_query' tool and provide your SQL-like statement. The agent will fetch the structural JSON blocks directly from Couchbase, allowing you to perform complex data retrieval using familiar SQL syntax.

**Can I search documents using full-text query logic?**
Absolutely. The 'fts_search' tool leverages Couchbase's Full-Text Search (FTS) engine. Provide an index name and a boolean query string to perform structural text-based extraction across your document trees.