Connect your account, then chat with AI to run tools.
List collections, inspect schemas, read and paginate documents, run aggregations, and insert or update records in Atlas or self-hosted MongoDB from your AI client.
Live probe refreshed Sep 14, 2026 · Endpoint host mcp.mcpbundles.com
Built for
App builders, operators, and agents working with document data in MongoDB
Explore an unfamiliar database
Uses list collections, inspect collection, and get documents.
List collections in my MongoDB database, inspect the largest collection, then show five sample documents and summarize the main fields.
Connect MongoDB to any MCP client in minutes
Try this workflow
Explore an unfamiliar database
List collections in my MongoDB database, inspect the largest collection, then show five sample documents and summarize the main fields.
You’ll sign in to MCPBundles when your client connects.
https://mcp.mcpbundles.com/bundle/mongodbThis server needs a sign-in. MCPBundles handles that setup once — then chat here with saved access and a log of what ran.
Opens MCPBundles Studio with this server selected. After sign-in, chat and run tools from the same thread.
Browse all toolsFind widgets by store
Typical app-data read with query, sort, skip, and limit.
In the widgets collection, list documents where storeId matches a given id, sorted by updated date, with skip and limit for pagination.
What is MongoDB MCP?
It connects your AI assistant to a MongoDB database you already use — Atlas, DocumentDB, or a local instance via MCPBundles Desktop — so the agent can list collections, read documents, run aggregations, and apply inserts or updates with your permission.
Do I need separate username and password fields?
Not when you connect with a full mongodb:// or mongodb+srv:// connection string that already includes credentials. Add optional fields only if you build the URI from host and password separately.
Domain knowledge for MongoDB — workflow patterns, data models, and gotchas for your AI agent.
MongoDB stores data in databases → collections → documents. Documents are JSON-like with _id. Works with Atlas, self-hosted, DocumentDB, Cosmos DB MongoDB API. Connection string format: mongodb+srv:// or mongodb://.
| Setup | Fields | Routing |
|---|---|---|
| Atlas / remote MongoDB | connection string (preferred) or host + database + optional user/password | Direct MongoDB wire protocol |
| Local MongoDB on this machine | database, user, password, port (+ desktop tunnel) | MCPBundles Desktop to localhost |
Mongodb Aggregate
Run a read-only aggregation pipeline on one collection ($match, $group, $lookup, etc.). Pass pipeline as a JSON array. Results are capped for safety.
Mongodb Create Collection
Create an empty collection in the current database.
Mongodb Create Index
Create an index on a collection. Pass keys as JSON with field names and 1 (asc) or -1 (desc).
Mongodb Delete Documents
Delete documents in a collection that match a filter. Query must not be empty {} — narrow the filter so you only remove intended records.
Mongodb Drop Collection
Drop a collection and all of its documents. Requires confirm=true.
Mongodb Drop Index
Drop a named index. Cannot drop the default _id index.
Mongodb Explain Operation
Return MongoDB explain output for a find filter or aggregation pipeline. Use to debug slow queries and index usage before running large reads.
Mongodb Export Collection
Export collection documents to JSON or JSONL. Small exports return inline data; larger exports spill to a presigned S3 download_url (over ~512 KB). Pass that download_url to mongodb_import_documents as file_url to re-import elsewhere.
Mongodb Get AI Description
Get a compact, token-efficient schema description optimized for AI text-to-query generation
Mongodb Get Documents
Read documents from one MongoDB collection. Pass document_id for a single record by _id (use the _id string from a prior list result). Omit document_id to list with query, limit, skip, and optional sort. Always set collection — discover names with list collections first.
Mongodb Get Server Info
Get MongoDB server version and build information
Mongodb Import Documents
Bulk-insert documents from a file. Use file_content_base64 for payloads up to 1 MB, file_url for larger HTTP(S) sources (including export spillover links), or local_path with MCPBundles Desktop for files on this computer (https://www.mcpbundles.com/desktop). Format: json array or jsonl.
Mongodb Insert Documents
Insert one or more documents into a MongoDB collection. Pass an array of JSON objects; _id may be omitted for server-generated ObjectIds. Use inspect collection or get documents to learn field shapes before inserting.
Mongodb Inspect Collection
Inspect one collection: estimated document count, index names, and a few sample documents to learn field shapes before writing filters for get documents.
Mongodb List Collections
List all collections in the current database
Mongodb List Databases
List all databases in the MongoDB instance with size information
Mongodb List Indexes
List index definitions for one collection (keys, uniqueness, TTL, partial filters). Use before create index or drop index.
Mongodb Replace Document
Replace one document by _id with a full new document body. Unlike update documents, this overwrites the entire document. Set upsert to create the document when missing.
Mongodb Search Collections
Find documents across multiple collections in one database. Applies the same filter to up to 20 collections (cap with max_collections). Each hit includes collection name. Prefer mongodb_get_documents when you already know the collection.
Mongodb Update Documents
Update all documents in a collection that match a filter. Pass a non-empty query and an update JSON object with operators such as $set. Set upsert true to insert when nothing matches.
It connects your AI assistant to a MongoDB database you already use — Atlas, DocumentDB, or a local instance via MCPBundles Desktop — so the agent can list collections, read documents, run aggregations, and apply inserts or updates with your permission.
Not when you connect with a full mongodb:// or mongodb+srv:// connection string that already includes credentials. Add optional fields only if you build the URI from host and password separately.
Add the MCPBundles server URL to your MCP client configuration (Claude Desktop, Cursor, VS Code, etc.). The URL format is: https://mcp.mcpbundles.com/bundle/mongodb. Authentication is handled automatically.
MongoDB provides 20 tools that can be called by AI agents, along with a SKILL.md that gives your AI agent domain knowledge about when and how to use them.
MongoDB uses API Key. MongoDB requires credentials. Connect via MCPBundles and authentication is handled automatically.
© 2026 ThinkChain Inc. All rights reserved.
MCPBundles is an independent platform built on the open Model Context Protocol standard. Not affiliated with Anthropic PBC or Claude.
Other MCP servers in this category from the directory index
Opinionated typography toolkit on top of Google Fonts. Mood search, heading/body pairing, modular ty...
5 toolsCreate technical diagrams using AI. Deliver consistent, accurate designs faster.
Base44 is a no-code AI development platform that turns builders’ ideas into fully functional apps an...
Workable is your system of record for everything HR. Find, hire, and manage top talent effortlessly....
Harness is a unified AI software delivery platform to manage the SDLC using purpose-built AI agents.