> 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/sdks/mcp-registry.md).

# Install from the MCP Registry

The Algenta MCP server exposes the whole engine — governed data, decisions, simulations, agent runs, and repository intelligence — as Model Context Protocol tools your AI client can call. It is published to the **Official MCP Registry** as `io.github.thyn-ai/algenta`, so registry-aware clients and directories can discover and install it by name. Under the hood it runs the `algenta-mcp` package and authenticates to your deployment with `ALGENTA_API_KEY`.

This page connects it to the most common clients. For the tool catalog and transports, see [MCP server](/sdks/mcp.md); for the tools grouped by capability, see [MCP tools overview](/mcp-tools/overview.md).

## Before you start

* An Algenta API key exported as `ALGENTA_API_KEY`. Get one from [Authentication & API keys](/getting-started/authentication.md) or run [`algenta login`](/getting-started/login.md).
* Your base URL: `https://api.algenta.ai` (hosted) or `http://localhost:8000` (self-hosted). Set it as `ALGENTA_BASE_URL`.
* The server package on your PATH:

```bash
pipx install algenta-mcp    # or: pip install algenta-mcp
```

{% hint style="info" %}
`algenta-mcp` runs the **stdio** transport by default — the right choice for the desktop clients below. The canonical environment variables are `ALGENTA_API_KEY` and `ALGENTA_BASE_URL` (legacy `DE_API_KEY` / `DE_BASE_URL` are still accepted).
{% endhint %}

## Connect your client

{% tabs %}
{% tab title="Cursor" %}
Add the server to `~/.cursor/mcp.json`:

```json
{
  "mcpServers": {
    "algenta": {
      "command": "algenta-mcp",
      "env": {
        "ALGENTA_API_KEY": "$ALGENTA_API_KEY",
        "ALGENTA_BASE_URL": "https://api.algenta.ai"
      }
    }
  }
}
```

Reload Cursor, then reference the server in chat: *"@algenta simulate this business case…"*.
{% endtab %}

{% tab title="Claude Desktop" %}
Add the server to `claude_desktop_config.json` (on macOS: `~/Library/Application Support/Claude/claude_desktop_config.json`):

```json
{
  "mcpServers": {
    "algenta": {
      "command": "algenta-mcp",
      "env": {
        "ALGENTA_API_KEY": "$ALGENTA_API_KEY",
        "ALGENTA_BASE_URL": "https://api.algenta.ai"
      }
    }
  }
}
```

Restart Claude Desktop. `algenta` appears in the tools panel.
{% endtab %}

{% tab title="Claude Code" %}
Add the server with the Claude Code CLI (`-e` passes environment variables; everything after `--` is the launch command):

```bash
claude mcp add algenta \
  -e ALGENTA_API_KEY=$ALGENTA_API_KEY \
  -e ALGENTA_BASE_URL=https://api.algenta.ai \
  -- algenta-mcp
```

Verify it registered:

```bash
claude mcp list
```

{% endtab %}

{% tab title="Codex" %}
Add the server to `~/.codex/config.toml`:

```toml
[mcp_servers.algenta]
command = "algenta-mcp"
env = { ALGENTA_API_KEY = "$ALGENTA_API_KEY", ALGENTA_BASE_URL = "https://api.algenta.ai" }
```

Restart Codex so it picks up the new server.

For repo-specific Codex workflows, keep durable project instructions in `AGENTS.md` and keep MCP credentials in `~/.codex/config.toml` or a trusted project `.codex/config.toml`. Codex reads the `AGENTS.md` instruction chain when it starts a run, and the same MCP config is shared by the CLI and IDE extension. The `algenta-mcp` command above is a stdio server; if you expose an HTTP/SSE MCP endpoint instead, configure it as a streamable HTTP server and keep the bearer token in an environment variable rather than inline text.
{% endtab %}
{% endtabs %}

{% hint style="warning" %}
On a **self-hosted** or **air-gapped** deployment, point `ALGENTA_BASE_URL` at your own engine (for example `http://localhost:8000`). Private profiles fail closed and never fall back to the Algenta cloud.
{% endhint %}

## Install via Smithery

[Smithery](https://smithery.ai) builds and runs the server from the repository's `smithery.yaml`. Install it into a client with the Smithery CLI:

```bash
npx -y @smithery/cli install algenta --client cursor
```

Swap `--client` for `claude`, `claude-code`, or another supported target. Smithery reads `smithery.yaml` for the run command and prompts for `ALGENTA_API_KEY` and `ALGENTA_BASE_URL`.

## Verify the connection

{% hint style="success" %}
**Expected result** — your client lists the Algenta tools (for example `get_contract`, `simulate`, `list_data`, `triage_repository`). Ask it to *"call get\_contract and summarize the available Algenta tools"* — a successful reply confirms the key and base URL are wired correctly.
{% endhint %}

## Discoverable everywhere

Because Algenta is in the Official MCP Registry, third-party directories index it automatically — no extra publishing step:

* **Glama** — mcp server catalog and quality scoring.
* **PulseMCP** — server discovery and changelogs.
* **mcp.so** — searchable MCP server index.

Searching any of them for "Algenta" surfaces the same `io.github.thyn-ai/algenta` server.

## Troubleshooting

<details>

<summary>The client shows no Algenta tools</summary>

Confirm `algenta-mcp` is on your PATH (`pipx install algenta-mcp`) and that the config file is valid JSON/TOML, then fully restart the client. Desktop clients only load MCP config at startup.

</details>

<details>

<summary>Tool calls fail with 401</summary>

`ALGENTA_API_KEY` is missing or invalid. Set it in the client's `env` block (or shell) and confirm it matches an active key — see [Authentication & API keys](/getting-started/authentication.md).

</details>

<details>

<summary>It cannot reach the engine</summary>

Check `ALGENTA_BASE_URL`. Use `https://api.algenta.ai` for hosted, or your own origin for self-hosted. The engine must be reachable from the machine running the client.

</details>

## 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>Browse the MCP tools</strong></td><td>Every tool, grouped by capability.</td><td><a href="/pages/ZHmoXvADTGBmIZthNjE0">/pages/ZHmoXvADTGBmIZthNjE0</a></td></tr><tr><td><strong>Read the MCP server guide</strong></td><td>Transports, HTTP mode, and options.</td><td><a href="/pages/9l598Hwbo4aGEvKi96R8">/pages/9l598Hwbo4aGEvKi96R8</a></td></tr><tr><td><strong>Get an API key</strong></td><td>Mint the key your client will use.</td><td><a href="/pages/UvGGKnKB6EqDONoxZRyB">/pages/UvGGKnKB6EqDONoxZRyB</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 dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.algenta.ai/sdks/mcp-registry.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

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.