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

# MCP Tools Overview

> Canal 3 — Connect AI agents to your product catalog with 22 MCP tools across 6 categories.

## What is Canal 3?

Canal 3 is Alana's MCP (Model Context Protocol) server. It exposes your product catalog as a set of tools that AI agents — Claude, GPT-4, Cursor, Windsurf, and any MCP-compatible client — can call directly during a conversation or workflow.

Unlike the REST API (Canal 2), MCP tools are designed for **agent consumption**: structured inputs, structured outputs, no pagination boilerplate.

## 22 Tools Across 6 Categories

| Category          | Tools | Description                         |
| ----------------- | ----- | ----------------------------------- |
| **Catalog**       | 4     | Search and browse products          |
| **Optimization**  | 4     | AI-powered content optimization     |
| **Search**        | 3     | Advanced search and recommendations |
| **Media/Context** | 3     | Image and brand context tools       |
| **Analytics**     | 3     | Performance and query analytics     |
| **Feed**          | 3     | Feed generation and status          |

See the individual category pages for full parameter references.

## Transport

Canal 3 uses **SSE (Server-Sent Events)** transport over HTTPS:

```
https://app.alana.shopping/api/mcp/sse
```

Most MCP clients default to SSE. No WebSocket configuration required.

## When to Use MCP vs REST

| Use case                                          | Recommended            |
| ------------------------------------------------- | ---------------------- |
| AI agent needs product data during a conversation | **MCP (Canal 3)**      |
| Scheduled batch sync to an external platform      | **Feed API (Canal 1)** |
| Custom dashboard or app integration               | **REST API (Canal 2)** |
| Platform-native product feed (Google, Meta)       | **Feed API (Canal 1)** |

## Quick Start

1. Get an API key from **Settings → API Keys** with `mcp:read` scope
2. Configure your MCP client with the SSE endpoint and Bearer token
3. Call tools like `search_products` or `optimize_product` from your agent

See [Connecting Agents](/en/api-reference/mcp/connecting-agents) for client-specific setup instructions.

## Tool Filtering

Running all 22 tools adds context to every agent conversation. Use environment variables to load only the tools your use case needs:

```bash theme={null}
ALANA_MCP_GROUPS=catalog,search    # Load only catalog + search tools
ALANA_MCP_TOOLS=search_products    # Load only a specific tool
```

See [Tool Filtering](/en/api-reference/mcp/tool-filtering) for full details.