• RETRIEVAL
  • CONTEXT
  • ANALYSIS
  • HUMAN

How to run vector similarity searches with your AI

Connect your agent to MongoDB Atlas once. After that, your AI finds relevant data by matching the meaning of your questions to your stored embeddings.

Ask AI about this page

Short answer

How can my AI perform vector similarity searches?

Your agent uses the search tool to query MongoDB Atlas collections by comparing vector embeddings. This lets your AI pull documents based on semantic intent rather than exact keyword matches. You get contextually relevant data for your queries.

The outcomes.

What your AI delivers.

The AI handles the math and the database calls.

RETRIEVAL

Semantic data lookup

The search tool finds documents that match the intent of your query. Your AI identifies relevant context even when words don't match exactly.

CONTEXT

Rich prompt grounding

The agent pulls specific document snippets from Atlas. This gives your AI the exact information it needs to answer complex questions.

ANALYSIS

Pattern recognition

The AI compares retrieved vectors to find clusters of similar information. You get a clear view of how your data relates to your questions.

HUMAN

Manual index management

The AI cannot modify your Atlas index or change vector dimensions. You must manage your database schema and indexing settings yourself.

The workflow

What your AI does when a query arrives.

The AI translates your natural language into a mathematical search.

  1. Vectorize the query

    Your agent converts your plain text question into a high-dimensional vector embedding.

    search

  2. Execute similarity search

    The AI uses the search tool to find the closest matching vectors within your Atlas collection.

    search

  3. Filter and rank

    The agent evaluates the returned documents to ensure they meet your specific criteria.

    search

  4. Synthesize answer

    The AI reads the retrieved data to provide a direct response to your original question.

    search

Try it

Copy these to start.

Use these prompts to trigger a search immediately.

Starting points

Replace the bracketed text with your actual collection names or specific search topics.

MongoDB Atlas Vector Search Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_rMRfujNKzotBfoPOFvzesaIZTdVkigbmunZrRwce/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — MongoDB Atlas Vector Search capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "mongodb-atlas-vector-search-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_rMRfujNKzotBfoPOFvzesaIZTdVkigbmunZrRwce/mcp"
    }
  }
}
Copy into chat04

  • Search my 'customer_feedback' collection for any complaints regarding shipping delays.

  • Find documents in the 'product_catalog' that are semantically similar to a waterproof hiking boot.

  • Using the vector search, find all support tickets related to account security issues.

  • Look through the 'research_papers' collection for entries discussing renewable energy trends.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect MongoDB Atlas Vector Search once, then ask.

Just provide your connection details. Your credentials stay encrypted, and you can start querying your Atlas data through your AI client instantly.

Connect MongoDB Atlas Vector Search to your AI

FAQ

How this task behaves

  • 01

    Can the AI delete documents from my Atlas collection?

    No. The AI only has permission to perform searches through the provided tools.

  • 02

    Does the AI create the vector indexes for me?

    No. You must set up your Atlas Vector Search indexes in MongoDB before the AI can query them.

  • 03

    How does the AI find similar items?

    It uses the search tool to perform a vector similarity query against your existing embeddings.

  • 04

    Can I use this with Claude?

    Yes, any MCP-compatible client like Claude can use this connector once it is connected.

  • 05

    Will the AI change my database schema?

    No. The AI's capabilities are limited to searching; it cannot alter your database structure.

  • More questions about MongoDB Atlas Vector Search? The Connector page answers them. See everything the MongoDB Atlas Vector Search Connector can do

Connect MongoDB Atlas Vector Search to Claude, Cursor, ChatGPT & more