Claude MongoDB Atlas connectorWhat Claude can do in your MongoDB Atlas account.

Runs a MongoDB aggregation pipeline on a collection: a chain of stages that filter, group and compute, so you get a summary table instead of a pile of raw documents.

Runs the same kind of pipeline across a whole database rather than a single collection, for questions that span more than one collection.

Returns Atlas Performance Advisor recommendations for a cluster: suggested indexes, indexes worth dropping, slow query logs and schema advice.

Shows which IP addresses and network ranges are allowed to reach an Atlas cluster, so you know exactly who can knock on the door.

Returns details on one specific Atlas cluster, including its state. MongoDB also uses it to track a cluster until it's available again after a pause.

Pulls the list of alerts for an Atlas project, so you can see at a glance what fired and on which cluster. Handy for a quick status check.

Returns the list of Atlas clusters, meaning the groups of servers that host your databases, so you know what's running and where. It's the natural starting point before inspecting one.

Displays the database users defined in Atlas, the accounts that apps and people use to connect to your data. Useful when access hygiene is on the agenda.

Lists the Atlas organizations your identity can reach, the top level of MongoDB's hierarchy. Useful as soon as one login spans several companies or business units.

Returns Atlas projects, the folders that group clusters, users and access rules inside an organization. It helps you pin down the right scope before any deeper question.

Describes the parts of Atlas Stream Processing, the service for processing data continuously: workspaces, processors, public or private connections. Nothing is started or stopped, it only describes.

Details a collection's indexes, including vector and full-text search indexes, the structures that make queries faster. A missing index often explains a slow page.

Describes a collection's schema, meaning the usual shape of its documents and their fields. Claude understands the data before querying it.

Returns a collection's size in megabytes, so you can spot what's taking up the most room on a cluster. Good for spotting which collections have grown far beyond the rest.

Returns the number of documents in a collection. A plain question, a single figure in return. No documents come back, only the total.

Returns statistics for a database: volume, object count and other health indicators worth keeping an eye on. Think of it as a health check for one database at a time.

Shows how MongoDB's query optimizer runs a query: the winning plan and its execution statistics. You see why a query is slow.

Runs a standard query against a collection and returns the documents that match your criteria. It's MongoDB's basic query, labeled as a read: no document is changed along the way.

Lists the collections in a database, MongoDB's equivalent of tables, so you know where to look. Each collection is a group of related documents.

Returns every database available through the current connection. From there, Claude knows which database to dig into next. The list depends on the connection opened to the cluster by the connect tool.

Opens a connection to an Atlas cluster from MongoDB's remote server. The database tools then rely on that connection.