Agentic AI and MPC are coming on strong in the database market. After the likes of Postgres and Oracle, now it's MongoDB's turn to incorporate MCP.

We've been examining the latest trend of MCP being utilized in order to provide AI capabilities to DBMs products. One such article was "Agentic AI For PostgreSQL" which was about Crystal DBA's Postgres MCP Pro server and Xata's Agent.
With those tools you can now tap into a pool of never-seen-before functionality, such as using your AI client like Claude to ask for diagnostics about your database in natural language using prompts such as:

• "Check the health of my database and identify any issues"
  
  
• "What are the slowest queries in my database? And how can I speed them up?"
  
  
• "My app is slow. How can I make it faster?"
  
  
• "Analyze my database workload and suggest indexes to improve performance"

Oracle has also added MCP support to two of its flagship tools, SQLcl and SQL Developer for VS Code, in an attempt to simplify user interaction with the database. This means that, instead of writing SQL, devs can now make requests such as:

 "create me a table with vector search"

while business users can now do reporting by asking:

"what were the most popular shoes being sold leading up to the Boston Marathon?".

Another big name dbms, MongoDB, now joins the bandwagon, with the release of its official MCP server. 

So what can you do with it?

First of all as in the other cases described, now you can hook up your AI coding assistant onto your MongoDB instance. Per the official statement:

The MongoDB MCP Server enables developer tools with MCP clients to interact directly with a MongoDB database and to handle a range of administrative tasks, such as managing cluster resources, as well as data-related operations like querying and indexing.

One particularly useful feature is that of having direct access to the schema in natural language, so that you can:

• Explore your MongoDB data; explain your database schema and ask questions about your data and its relationships.
  
  
• Query your data; run CRUD operations in natural language and perform complex aggregations on your data without writing any code.

Then, Administrative-wise you can, again in natural language, manage you MongoDB deployments, including creating and managing databases, collections, and indexes.

That goes for your Atlas clusters too; you can create and manage Atlas organizations, projects, and clusters, and customize access to your data.

All that is possible thanks to the rich variety of tools the MCP server exposes that get coordinated by the LLM to perform the task at hand. A few of them are:

• atlas-connect-cluster - Connect to MongoDB Atlas cluster
• atlas-create-access-list - Allow Ip/CIDR ranges to access your MongoDB Atlas clusters.
• atlas-create-db-user - Create an MongoDB Atlas database user
• atlas-create-free-cluster - Create a free MongoDB Atlas cluster
• aggregate - Run an aggregation against a MongoDB collection
• collection-indexes - Describe the indexes for a collection
• collection-schema - Describe the schema for a collection
• collection-storage-size - Gets the size of the collection

The easiest way to get started with the MCP server is to get hold of the Docker provided image and have your MongoDB connection string or Atlas API credentials ready since the server will not start unless configured.

You then have to go through the usual drill of configuring your MCP clients, like Claude desktop or an alternative. Code examples and more instructions are on the project's Github repository.

More Information

Announcing the MongoDB MCP Server

Related Articles

Agentic AI For PostgreSQL 

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on Twitter, Facebook or Linkedin.

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info