# Chat and Build with AI

We now have an MCP server that gives any AI coding assistant direct access to our documentation

### What it does

Three tools are available to any connected AI client:

* **search\_docs** - Search our docs by keyword
* **get\_page** - Get full page content in markdown
* **list\_pages** - Browse all pages and their hierarchy

### Why this matters

Instead of copy pasting docs into prompts or hoping the AI has seen our documentation, you can now ask your AI assistant questions about UR directly and it will search and read the actual docs to answer. This works for coding, writing specs, onboarding, or anything where you need accurate information about our product.

### How to set it up

#### Claude Desktop

Add to your config file:

* macOS: `~/Library/Application Support/Claude/claude_desktop_config.json`
* Windows: `%APPDATA%\Claude\claude_desktop_config.json`

```json
{"mcpServers": {"ur-docs": {"url": "https://ur-docs-mcp-production.up.railway.app/mcp"}}}
```

#### Claude Code

Run this in your terminal:

```bash
claude mcp add ur-docs --transport http https://ur-docs-mcp-production.up.railway.app/mcp
```

#### Cursor

Open **Settings > MCP Servers > Add new server**:

* Name: `ur-docs`
* Type: `HTTP`
* URL: `https://ur-docs-mcp-production.up.railway.app/mcp`

#### VS Code (Copilot)

Add to `.vscode/mcp.json` in your project:

```json
{"servers": {"ur-docs": {"url": "https://ur-docs-mcp-production.up.railway.app/mcp"}}}
```

#### Windsurf

Add to `~/.codeium/windsurf/mcp_config.json`:

```json
{"mcpServers": {"ur-docs": {"serverUrl": "https://ur-docs-mcp-production.up.railway.app/mcp"}}}
```

### Try it

After setup, restart your AI client and ask:

* "What is UR?"
* "How does the user lifecycle work?"
* "What integration methods are available?"
* "Explain the core banking architecture"
* "Search UR docs for KYC"

The assistant will search and read our docs to give you an accurate answer.