• INDEXED
  • REMOVED
  • SYNCED
  • HUMAN

How to update RAG knowledge bases with your AI

Connect Typesense Vector Search once. From then on, your AI manages the data flow between your JSON sources and your vector collections.

Ask AI about this page

Short answer

How can I update my RAG knowledge base using AI?

Your AI reads your JSON data and uses the Typesense Vector Search MCP to index that information into searchable vector documents. This keeps your RAG retrieval accurate and up to date. You'll see how the AI handles indexing and deletions below.

Data destinations

Where your data lands.

The AI moves data between your raw JSON and your Typesense collections.

INDEXED

New vector documents

The AI converts JSON objects into searchable vectors. These live in your Typesense collections for immediate RAG use.

REMOVED

Cleaned collections

When data becomes obsolete, the AI removes specific documents. This prevents your RAG from retrieving stale or incorrect information.

SYNCED

Updated knowledge

Your AI keeps the vector store in line with your source files. This ensures your agent always has the latest context.

HUMAN

Schema decisions

The AI cannot change your Typesense collection schema. You must define your field types and vector dimensions yourself.

The workflow

What your AI does when JSON data arrives.

The AI acts as the bridge between your raw data and your search engine.

  1. Parse JSON

    Your AI reads the raw JSON files or streams you provide. It identifies the text and metadata needed for embedding.

    index_document

  2. Create vectors

    The AI prepares the data for the Typesense collection. It uses the index tool to push the new documents into the vector store.

    index_document

  3. Clean up old data

    If you tell the AI that a record is no longer valid, it finds the specific document in Typesense and removes it.

    delete_document

  4. Verify state

    The AI confirms the document was successfully added or removed. This ensures your RAG knowledge base is actually updated.

    index_documentdelete_document

Try it

Copy these to start.

Use these prompts to trigger the indexing process.

Starting points

These are starting lines. Swap out the bracketed text with your actual file names or collection names.

Typesense Vector Search Connector

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

Connector linkhttps://edge.vinkius.com/vk_preview_bmlNohSGzzJmxN6goNUndx2uYrEvS88xXUYpe92K/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 — Typesense Vector Search capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "typesense-vector-search-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_bmlNohSGzzJmxN6goNUndx2uYrEvS88xXUYpe92K/mcp"
    }
  }
}
Copy into chat04

  • Index this JSON file into my 'products' collection: [file_path]

  • Remove the document with ID [id] from my Typesense collection.

  • Update my RAG knowledge base by indexing the data in [file_name].

  • Delete all outdated entries in the 'user_logs' collection based on this JSON list.

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

Start here

Connect Typesense Vector Search once, then ask.

One click connects your AI to the MCP. Your credentials stay encrypted, and you can start indexing data immediately.

Connect Typesense Vector Search to your AI

FAQ

How this task behaves.

  • 01

    Can the AI create new Typesense collections?

    No. The AI can only index or delete documents within existing collections you have already created.

  • 02

    Does the AI handle the vector embedding process?

    Yes. The AI prepares the data so that the index tool can place it into the vector space correctly.

  • 03

    Will the AI delete my entire collection?

    No. The AI only uses the tool to remove specific documents you identify. It cannot drop or wipe a whole collection.

  • 04

    Can I use this to update a database other than Typesense?

    No. This specific MCP is built only for Typesense collections.

  • 05

    Does the AI need my Typesense API key?

    Yes, but you provide it through the Vinkius connection, which keeps it encrypted and secure.

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

Connect Typesense Vector Search to Claude, Cursor, ChatGPT & more