# DesiData [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/krishnakaushik195/desidata  
**GitHub Stars:** 0  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/desidata

## Description
Search India datasets; inspect sources, licences and previews; get notebook and Python examples.

## Claude Desktop Quick Installation
No confirmed install details yet, so no config is published. Follow the setup steps in the repository above.

## Documentation

## What DesiData MCP server does

The DesiData MCP server gives AI assistants an interface to the DesiData dataset catalogue. It is intended for finding India-focused datasets and gathering enough context to decide whether a dataset fits a research, analysis, or development task.

The server can return dataset metadata, provenance details, and a small preview of dataset contents. It can also provide examples showing how to load a dataset from a notebook or Python program. These capabilities help an agent move from discovery to an initial implementation without requiring the user to search the DesiData website manually.

## How it works

An assistant sends requests through the MCP connection to search the published catalogue or inspect a selected dataset. Search and preview requests are read-only. The returned information can include source and licence details, dataset metadata, sample data, and code-oriented loading guidance.

The DesiData MCP server does not replace the regular dataset download flow. Full downloads continue to use the user’s own DesiData account token through the normal DesiData access process. The MCP plugin therefore supports catalogue exploration and dataset evaluation, while authenticated downloads remain separate.

## Setup and configuration

The repository provides a remote MCP service for the plugin. The README identifies Cursor, Codex, and Gemini CLI as supported clients. For Gemini CLI, install the repository extension with:

```sh
gemini extensions install https://github.com/krishnakaushik195/desidata
```

Restart Gemini CLI after installation and run `/mcp list` to verify that the connection is available. Other clients should follow the connection instructions in the plugin’s setup guide.

No account or API token is needed for catalogue searches or previews. If a workflow proceeds to a full dataset download, create a free DD token from the DesiData profile and provide it through the normal download flow. Keep `DD_TOKEN` private; do not commit it to a notebook or repository.

## Tools and capabilities

The DesiData MCP server supports these documented operations:

- Search the published DesiData catalogue.
- Inspect dataset metadata.
- Review dataset sources and provenance.
- Check licence information.
- Preview a small sample of a dataset.
- Retrieve notebook examples.
- Retrieve Python loader examples.

The examples are useful when an assistant needs to explain how to work with a selected dataset in Google Colab or another Python environment.

## Limitations and notes

The documented MCP capabilities cover search, inspection, previews, and code examples; the README does not describe an MCP tool for downloading complete datasets directly. Full downloads use the user’s own `DD_TOKEN` through DesiData’s regular authenticated flow.

Some repository notebook download cells may require updates because the download process now uses the authenticated token flow. Preview results are samples rather than confirmation that a complete file has been downloaded or processed. The remote service’s current limits and connection details are maintained in the plugin setup guide.

_Full upstream README: https://allmcps.com/mcp/desidata/readme_