mongodb-mcp-server

Establishes and maintains persistent connections to MongoDB instances and MongoDB Atlas clusters through the Model Context Protocol, handling authentication (connection strings, credentials), connection pooling, and lifecycle management. Implements MCP server transport layer to expose MongoDB as a resource accessible to LLM clients and agents without direct database access.

Executes MongoDB queries (find, aggregate, insert, update, delete) through MCP tools that accept query parameters and return structured results. Implements query validation and schema introspection to provide type information about collections, enabling LLMs to construct valid queries without trial-and-error. Uses MongoDB's aggregation pipeline and query language natively.

Provides MCP tools to execute geospatial queries on MongoDB collections with 2dsphere or 2d indexes. Implements MongoDB's geospatial operators ($near, $geoWithin, $geoIntersects) enabling agents to find documents based on geographic proximity or containment. Supports GeoJSON format for location data.

Provides MCP tools to perform faceted search and analytics using MongoDB's aggregation framework. Agents can request facets (counts by category, range, etc.) alongside search results, and execute complex analytics queries that group, filter, and transform data. Implements multi-facet aggregation pipelines for exploratory data analysis.

Provides MCP tools to export MongoDB query results in multiple formats (JSON, CSV, BSON) and handle large result sets through pagination or streaming. Implements result formatting and serialization, enabling agents to extract data for external processing or reporting. Supports configurable field selection and transformation during export.

Provides MCP tools to list databases, collections, and indexes, and retrieve schema information including field names, types, and validation rules. Implements MongoDB's introspection APIs (listDatabases, listCollections, getIndexes) and potentially uses schema inference or validation metadata to expose structure to LLM clients. Enables agents to discover available data without prior knowledge of the database structure.

Provides a dedicated MCP tool for constructing and executing MongoDB aggregation pipelines, which are multi-stage data transformation workflows. Accepts pipeline stages (match, group, project, sort, limit, etc.) as structured input and executes them server-side, returning transformed results. Implements validation of pipeline syntax and stage compatibility before execution.

Provides MCP tools for inserting single documents, inserting multiple documents in bulk, and performing bulk write operations (mixed insert/update/delete). Implements validation of document structure before insertion and handles MongoDB's write concern and error handling. Supports ordered and unordered bulk operations with configurable behavior on partial failures.

Provides MCP tools for updating documents (updateOne, updateMany) and deleting documents (deleteOne, deleteMany) using MongoDB query filters and update operators. Implements MongoDB's update operators ($set, $inc, $push, etc.) and supports both replacement and partial updates. Handles write concern and returns operation results including matched and modified counts.

Provides MCP tools to list existing indexes on collections and potentially create or drop indexes. Implements MongoDB's index APIs (createIndex, dropIndex, getIndexes) and may expose query plan information to help agents understand index usage. Enables agents to optimize queries by understanding or creating appropriate indexes.

Provides MCP tools to execute MongoDB transactions, which ensure ACID guarantees across multiple documents and collections. Implements MongoDB's session-based transaction API (startSession, startTransaction, commitTransaction, abortTransaction) and handles transaction state management. Enables agents to perform complex multi-step operations with rollback capability.

Provides MCP tools to establish change streams on collections or databases, enabling agents to monitor real-time data modifications. Implements MongoDB's change stream API with support for filtering change events and resuming from specific points. Agents can subscribe to insert, update, delete, and replace events and react to data changes in real-time.

Provides MCP tools to perform full-text search on MongoDB collections using text indexes and the $text query operator. Implements MongoDB's text search capabilities including language support, stemming, and relevance scoring. Agents can search across multiple fields and retrieve results ranked by relevance.