> ## Documentation Index
> Fetch the complete documentation index at: https://docs.soriaanalytics.com/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP

> Connect to Soria's MCP endpoint and use the available data tools.

# MCP

Soria's remote MCP endpoint is:

```text theme={null}
https://mcp.soriaanalytics.com
```

Install it using the [ChatGPT guide](/guides/chatgpt) or [Claude guide](/guides/claude). Both use your existing Soria account through browser OAuth.

## Tools

| Tool                      | What it does                                                               |
| ------------------------- | -------------------------------------------------------------------------- |
| `list_datasets`           | Lists datasets available to the signed-in account.                         |
| `search_datasets`         | Finds datasets by topic, domain, or column.                                |
| `get_dataset_info`        | Returns a dataset's description, grain, cadence, and providers.            |
| `get_dataset_schema`      | Returns columns, accepted values, methodology, and source lineage.         |
| `get_dataset_stats`       | Returns row count or a bounded profile of one column.                      |
| `query_dataset`           | Selects, filters, sorts, and paginates rows from one dataset.              |
| `search_documents`        | Searches the public source documents behind Soria datasets.                |
| `list_dataset_files`      | Lists source files contributing to a dataset.                              |
| `get_dataset_source_file` | Returns a temporary URL for downloading one verified original source file. |

A typical request starts with `search_datasets`, continues with `get_dataset_schema`, and then uses `query_dataset`. Use the source-file tools when you need the original evidence behind a dataset.

See [Download a source file](/guides/source-files) for the exact flow.