• MATCHES
  • DETAILS
  • CONTEXT
  • HUMAN

How to perform semantic search with your AI

Connect Vald to your agent to query massive, distributed datasets. Your agent can now find data based on meaning rather than just keyword matches.

Ask AI about this page

Short answer

How can I run semantic searches with my AI?

Your agent sends query vectors to Vald to find the closest matches within distributed datasets. This allows you to pull specific information based on conceptual similarity. You'll see the specific outcomes below.

Search outcomes

Where your data lands

These are the results your agent produces after a query.

MATCHES

Closest vector matches

The agent identifies the most relevant data points in the distributed set. You get a list of items that match the intent of your query.

DETAILS

Specific vector data

The agent pulls the underlying information tied to a specific vector. This gives you the actual content behind the match.

CONTEXT

Semantic relevance

The agent filters results based on how well they align with your question. You get answers that respect the meaning of your input.

HUMAN

Manual review

You look over the retrieved data to confirm it's what you actually need. The AI provides the matches, but you make the final call.

The workflow

What your AI does when a query arrives.

The agent handles the heavy lifting of searching through distributed vectors.

  1. Find similar vectors

    The agent processes your input to find the most relevant vectors across the distributed dataset.

    search_vectors

  2. Retrieve vector data

    The agent pulls the specific metadata or content associated with the identified vectors.

    get_vector_details

  3. Verify similarity

    The agent uses the search function again to refine the results based on your specific query intent.

    search_vectors

  4. Extract details

    The agent fetches the final set of information you need to answer your question.

    get_vector_details

Try it

Copy these to start.

Use these prompts to begin searching your data.

Starting points

These are opening lines. Replace the bracketed text with your specific dataset names or topics.

Vald Connector

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

Connector linkhttps://edge.vinkius.com/vk_preview_LwabvjFRxR6SLDMyq0Lg8wB1fanBWcNw0aRuqiDM/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 — Vald capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "vald-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_LwabvjFRxR6SLDMyq0Lg8wB1fanBWcNw0aRuqiDM/mcp"
    }
  }
}
Copy into chat04

  • Find the most relevant entries in the [dataset name] regarding [topic].

  • Search the [dataset name] for vectors similar to [specific concept].

  • Get the details for the vector closest to [query term] in the [dataset name].

  • What are the most similar data points in [dataset name] to the concept of [topic]?

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

Start here

Connect Vald once, then ask.

One link is all it takes. Your login stays encrypted on our side, and you can find the full walkthrough on the Connector page.

Connect Vald to your AI

FAQ

How semantic search behaves

  • 01

    Can the AI change or delete my vectors?

    No. The tools only allow the agent to search and retrieve data. It cannot modify or delete anything in your datasets.

  • 02

    Does the AI search my entire database at once?

    The agent searches the distributed datasets you have connected through the Vald MCP.

  • 03

    Can the AI create new vectors for me?

    No. There are no tools available to create or write new vectors to the store.

  • 04

    How is this different from a standard text search?

    It uses vector similarity to find data based on the meaning of your query, not just matching exact words.

  • 05

    Can the AI update the metadata of a vector?

    No. The agent can only retrieve existing details; it cannot update them.

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

Connect Vald to Claude, Cursor, ChatGPT & more