> For the complete documentation index, see [llms.txt](https://docs.algenta.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.algenta.ai/mcp-tools/overview.md).

# MCP tools overview

The Algenta MCP server exposes the engine as Model Context Protocol tools for Cursor, Claude, and other hosts. Browse the tool catalog by domain.

The Algenta MCP server (`algenta-mcp`) turns your engine into a set of Model Context Protocol tools. Point an MCP host at your deployment and your assistant can discover datasets, run governed queries, manage connectors, simulate decisions, and more — every tool a thin, authenticated wrapper over one `/v1` endpoint. This section is the reference for the tools themselves, grouped by domain. For installing and connecting the server, see the [MCP server](/sdks/mcp.md) page.

{% hint style="info" %}
Each MCP tool maps one-to-one to an Algenta API endpoint. Anything you can do through a tool you can reproduce with a direct `curl` against the same `/v1` path — useful for verifying behavior outside an MCP host.
{% endhint %}

## How the tools appear in your host

Once the server is configured, the tools show up in your host's tool panel after a restart, named exactly as in the reference tables below (for example `list_data`, `query_data`, `create_connector`). You drive them in natural language; the assistant picks the tool and fills the arguments.

| Host                                            | Transport       | Config                                       |
| ----------------------------------------------- | --------------- | -------------------------------------------- |
| Cursor, Claude Desktop, Zed, Windsurf, VS Code  | stdio           | Add `algenta-mcp` to the host's MCP config.  |
| ChatGPT, OpenWebUI, LibreChat, n8n, custom apps | Streamable HTTP | Point the client at `$ALGENTA_BASE_URL/mcp`. |

The tool schemas are always live. Enumerate the exact set your deployment exposes with:

```bash
curl -s $ALGENTA_BASE_URL/mcp/tools | jq -r '.tools[].name'
```

{% hint style="info" %}
The visible tool set depends on your key's entitlements and product edition. If a tool below is not in the live list for your deployment, your key is not entitled to it. Start any session with `get_contract` to read the machine-readable capability contract.
{% endhint %}

## Tool groups

Each page is a reference: a table of every tool in the group, the method and endpoint each maps to, and worked examples you can run from an MCP host or reproduce with `curl`.

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><strong>Onboard and ingest data</strong></td><td>Register and train datasets, run the high-level connection flow, register sources, and ingest tables to a simulation.</td><td><a href="/mcp-tools/data.md">Data &amp; ingest tools</a></td></tr><tr><td><strong>Manage saved connectors</strong></td><td>Create, inspect, test, browse, update, and delete reusable data connectors — plus preview inline definitions.</td><td><a href="/mcp-tools/connectors.md">Connector tools</a></td></tr><tr><td><strong>Run governed queries</strong></td><td>Governed exact queries, multi-metric batches, and constrained read-only SQL reports over your datasets.</td><td><a href="/mcp-tools/query.md">Query tools</a></td></tr></tbody></table>

{% hint style="info" %}
The server exposes many more tools than these three groups — LLM utilities, decisions and simulations, agent runs, triggers, repository intelligence, and control-plane operations among them. The [MCP server](/sdks/mcp.md) page catalogs the core set, and `get_contract` returns the full contract for your deployment.
{% endhint %}

## Discover, then act

The data and query tools are designed to be chained low-token first, then exact — cheap discovery before you pay for a full schema or a query.

{% stepper %}
{% step %}

### Read the contract

`get_contract` returns the machine-readable contract: tool names, endpoint families, and the governed-filter rules for this deployment.
{% endstep %}

{% step %}

### Find datasets cheaply

`list_data(search=..., compact=true)` for low-token dataset discovery, then `get_data_summary(dataset_id)` before loading the full schema.
{% endstep %}

{% step %}

### Load the schema only when needed

`get_data_schema(dataset_id)` returns columns, roles, formulas, and query hints — reach for it once you have chosen a dataset.
{% endstep %}

{% step %}

### Run governed queries

`query_data` for a single governed exact query, `query_batch` for several in one call, and `query_sql_report` for wide read-only rowsets.
{% endstep %}
{% endstepper %}

## Next steps

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><strong>Set up the MCP server</strong></td><td>Install, configure, and connect Algenta to Cursor, Claude Desktop, and HTTP hosts.</td><td><a href="/sdks/mcp.md">MCP server</a></td></tr><tr><td><strong>Discover tools and the contract</strong></td><td>Bootstrap any client from the live contract and tool list.</td><td><a href="/sdks/tool-discovery.md">Tool discovery</a></td></tr><tr><td><strong>Browse the HTTP API</strong></td><td>The one contract every tool wraps.</td><td><a href="/http-api/overview.md">API overview</a></td></tr></tbody></table>


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.algenta.ai/mcp-tools/overview.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.