MongoDB Integration

Connect to MongoDB databases and sync document collections to your knowledge base. Perfect for flexible, document-oriented data.

How It Works

The MongoDB connector reads documents from your collections and converts them into searchable text. Nested documents and arrays are automatically flattened for comprehensive indexing.

• Automatically discovers collections and fields
• Handles nested documents and arrays
• Supports selective collection and field sync
• Enables incremental sync via timestamp or ObjectId

Connecting MongoDB

mongodb+srv:500">class=500">class="text-green-500">"text-muted-foreground">//user:password@cluster.mongodb.net/database

Connection Settings

Sync Settings

Nested Documents

MongoDB documents often contain nested objects and arrays. The connector automatically flattens these for searchability:

{
  "_id": "507f1f77bcf86cd799439011",
  "title": "Product Guide",
  "author": {
    "name": "Jane Smith",
    "department": "Documentation"
  },
  "tags": ["guide", "getting-started", "tutorial"],
  "sections": [
    { "title": "Introduction", "content": "Welcome to..." },
    { "title": "Setup", "content": "First, install..." }
  ]
}

Becomes searchable text:

Collection: articles
title: Product Guide
author.name: Jane Smith
author.department: Documentation
tags: guide, getting-started, tutorial
sections: title: Introduction, content: Welcome to..., title: Setup, content: First, install...

Field Selection with Dot Notation

You can specify exactly which fields to index using dot notation for nested fields:

{
  "articles": ["title", "content", "author.name"],
  "products": ["name", "description", "specs.features"]
}

Incremental Sync

MongoDB supports two methods for incremental sync:

1. Timestamp Field

If your documents have an updatedAt or similar field:

• Only documents modified since the last sync are processed
• Great for documents that track modification time

2. ObjectId-Based

If you don't have a timestamp field, the connector can use MongoDB's _id field:

• Only documents created after the last sync are processed
• Doesn't catch updates to existing documents

MongoDB Atlas

For MongoDB Atlas (cloud) deployments:

1. Use the connection string from Atlas Dashboard → Connect → Connect your application
2. Enable TLS (required for Atlas)
3. Ensure your IP is allowlisted in Atlas Network Access settings

mongodb+srv:500">class=500">class="text-green-500">"text-muted-foreground">//username:password@cluster0.xxxxx.mongodb.net/myDatabase?retryWrites=500">true&w=majority

Security Best Practices

• Create a read-only user: Never use admin credentials
• Limit collection access: Grant read only on needed collections
• Use TLS: Always enable TLS for production databases
• IP allowlisting: Restrict database access to RAG Chats IPs
• Exclude sensitive data: Don't sync collections with PII

Example: Creating a Read-Only User

500">class=500">class="text-green-500">"text-muted-foreground">// Connect to admin database
use admin

500">class=500">class="text-green-500">"text-muted-foreground">// Create a read-only user 500">for specific database
db.createUser({
  user: 500">class="text-green-500">"ragchats_reader",
  pwd: 500">class="text-green-500">"secure_password",
  roles: [
    { role: 500">class="text-green-500">"read", db: 500">class="text-green-500">"myDatabase" }
  ]
})

500">class=500">class="text-green-500">"text-muted-foreground">// Or 500">for specific collections only
db.createUser({
  user: 500">class="text-green-500">"ragchats_reader",
  pwd: 500">class="text-green-500">"secure_password",
  roles: [
    {
      role: 500">class="text-green-500">"read",
      db: 500">class="text-green-500">"myDatabase",
      collection: 500">class="text-green-500">"articles"
    },
    {
      role: 500">class="text-green-500">"read",
      db: 500">class="text-green-500">"myDatabase",
      collection: 500">class="text-green-500">"products"
    }
  ]
})

Troubleshooting

Connection failed

• Verify connection string or host/port settings
• Check if MongoDB server is running and accessible
• Ensure firewall rules allow the connection
• For Atlas, verify IP is allowlisted
• Check TLS settings match server requirements

Authentication failed

• Verify username and password
• Check auth source database (usually "admin")
• Ensure the user exists and has correct permissions

No data synced

• Verify collections contain documents
• Check if collections are empty
• Ensure selected collections exist
• System collections (starting with "system.") are skipped

Nested data not appearing

• Increase max depth setting
• Check if nested fields are specified correctly
• Very deep nesting (10+) is truncated

Sync is slow

• Reduce the number of collections being synced
• Use incremental sync with a timestamp field
• Increase batch size for faster processing
• Add an index on your timestamp field

Removing the Integration

1. In RAG Chats, delete the MongoDB data source
2. Optionally, remove the database user:

   use admin
   db.dropUser(500">class="text-green-500">"ragchats_reader")