• SEARCH
  • FILTER
  • DATA
  • HUMAN

How to query and filter vector data with your AI

Connect Zilliz Cloud to your AI client to turn natural language questions into precise vector searches. Once connected, your agent handles the heavy lifting of finding relevant data points.

Ask AI about this page

Short answer

How can I search and filter my vector data using natural language?

Your AI uses similarity searches to find relevant vectors and applies metadata filters to narrow down results. You get specific, filtered data back by simply describing what you need in plain English. Here is how the process works.

The results you get

Where your data lands

The AI translates your intent into specific database operations.

SEARCH

Relevant vector matches

The AI finds the closest data points to your query. You receive the most semantically similar results from your collection.

FILTER

Narrowed results

The AI applies metadata constraints to your search. This ensures you only see data that meets your specific criteria.

DATA

Structured entity info

The AI pulls specific entity details based on your filters. You get the exact attributes you asked for.

HUMAN

Manual review

You decide if the returned vectors actually meet your needs. The AI provides the candidates, but you make the final call.

The workflow

What your AI does when the request arrives.

The AI acts as the bridge between your chat interface and the Zilliz Cloud database.

  1. Interpret intent

    Your agent reads your prompt to determine if you want a similarity search or a specific metadata lookup.

    search_vectorsquery_entities

  2. Find similar vectors

    The AI executes a search to find data points that are mathematically close to your input.

    search_vectors

  3. Apply filters

    The agent uses metadata to prune the results, ensuring the search stays within your required parameters.

    query_entities

  4. Format response

    The AI takes the raw data returned from the database and presents it as a readable answer.

    query_entities

Try it

Copy these to start.

Use these prompts to test the connection immediately.

Starting points

These are examples. Swap out the collection names and attribute keys to match your actual Zilliz Cloud setup.

Zilliz Cloud Connector

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

Connector linkhttps://edge.vinkius.com/vk_preview_RKv8SXkDaOvT7RS1nwfoO7ZP8GbUdouO1cKQtzMb/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 — Zilliz Cloud capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "zilliz-cloud-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_RKv8SXkDaOvT7RS1nwfoO7ZP8GbUdouO1cKQtzMb/mcp"
    }
  }
}
Copy into chat04

  • Find the most similar vectors in my 'product_embeddings' collection to the concept of 'sustainable packaging'.

  • Search for entities in the 'user_profiles' collection where the 'location' metadata is set to 'New York'.

  • Query the 'document_vectors' collection for topics related to 'climate change' but only for documents created in 2023.

  • Show me the top 5 most similar items to 'blue running shoes' in my 'inventory' collection.

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

Start here

Connect Zilliz Cloud once, then ask.

Follow the link to authorize the connection. Your credentials stay encrypted, and you can start querying immediately.

Connect Zilliz Cloud to your AI

FAQ

How the task behaves

  • 01

    Can the AI delete data from my Zilliz Cloud collection?

    No. The available tools only allow for searching and querying data, not deleting or modifying it.

  • 02

    Can I use this to update my vector embeddings?

    No. This connector is designed for retrieval and searching. It does not have tools to write or update vectors.

  • 03

    Does the AI handle the metadata filtering automatically?

    Yes. If you include a filter in your natural language prompt, the AI uses the query tools to apply those constraints.

  • 04

    Can I perform a search without any metadata?

    Yes. You can perform a pure similarity search using only the vector search capability.

  • 05

    Will the AI decide which vectors are the most important?

    The AI uses the mathematical similarity scores from the database to find matches, but you are the one who interprets the relevance of those matches.

  • More questions about Zilliz Cloud? The Connector page answers them. See everything the Zilliz Cloud Connector can do

Connect Zilliz Cloud to Claude, Cursor, ChatGPT & more