• INVENTORY
  • INSPECTION
  • AUDIT
  • HUMAN

How to manage Ragas datasets with your AI

Connect Ragas to your AI client to browse and inspect your evaluation data. Once linked, you can pull dataset details directly into your chat.

Ask AI about this page

Short answer

How can I manage Ragas datasets with my AI?

You can list all your available datasets and pull the specific contents of any individual set using your AI client. This lets you audit your evaluation data without leaving your chat window. You'll see the specific outcomes below.

Data outcomes.

Where your data goes.

Your AI handles the retrieval so you can focus on the evaluation results.

INVENTORY

Full dataset list

The AI scans your Ragas account to show you every dataset you've created. You get a clean list of names to work with.

INSPECTION

Dataset contents

The AI pulls the specific rows and test cases from a chosen dataset. You get the raw data right in your chat.

AUDIT

Data verification

Since the AI can read your datasets, you can ask it to check for specific patterns or missing fields. This makes spotting errors easy.

HUMAN

Manual management

The AI cannot edit or delete your data. You'll still need to use the Ragas interface for any structural changes or deletions.

The workflow

What your AI does when the request arrives.

The AI uses these specific tools to bridge your chat interface with your Ragas storage.

  1. Cataloging

    The AI identifies every dataset currently stored in your Ragas account.

    list_datasets

  2. Targeting

    The AI selects the exact dataset you want to inspect based on your request.

    get_dataset

  3. Retrieval

    The AI pulls the specific data points or test cases you asked for.

    get_dataset

  4. Presentation

    The AI formats the retrieved dataset information so it's easy to read in your client.

    get_dataset

Try it

Copy these to start.

Paste these into your chat to begin interacting with Ragas.

Starting points

Swap the dataset names in these examples for the ones you actually use.

Ragas Connector

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

Connector linkhttps://edge.vinkius.com/vk_preview_mQUV7klN4rh09rRcp9ld3cwceCP25oL5Lk8res4B/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 — Ragas capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "ragas-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_mQUV7klN4rh09rRcp9ld3cwceCP25oL5Lk8res4B/mcp"
    }
  }
}
Copy into chat04

  • List all my Ragas datasets.

  • Show me the contents of the 'eval_v1' dataset.

  • Get the details for my 'test_cases_final' dataset.

  • What datasets are currently in my Ragas account?

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

Start here

Connect Ragas once, then ask.

Link your account through the Vinkius dashboard. Your credentials stay encrypted, and you'll be ready to chat with your datasets immediately.

Connect Ragas to your AI

FAQ

How this task behaves.

  • 01

    Can the AI delete my datasets?

    No. The AI can only list and retrieve data. You must use the Ragas UI to delete anything.

  • 02

    Can I add new data to a dataset via chat?

    No. This connector is read-only. It can list and get data, but it cannot create or modify it.

  • 03

    Can the AI see all my datasets?

    Yes, it can see all datasets when you ask it to list them, but it only pulls the details for the specific one you request.

  • 04

    How do I see the actual test cases?

    Ask your AI to get the contents of a specific dataset by name.

  • 05

    Does this work with any AI client?

    Yes, it works with any MCP-compatible client like Claude, Cursor, or Windsurf.

  • More questions about Ragas? See everything the Ragas Connector can do

Connect Ragas to Claude, Cursor, ChatGPT & more