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

# Installation

> Installation instructions how to connect Delora to your AI environment.

You can either run the MCP server locally using the official Delora MCP npm package or connect directly to the hosted MCP endpoint.

## 1. Run the Delora MCP Package Locally

Run the Delora MCP package locally.

### Quick Start (npx)

Run directly without installing globally:

```bash theme={null}
npx -y @deloraprotocol/mcp@latest
```

<Info>
  This starts the server in stdio mode (default), which is suitable for most local AI agent integrations.
</Info>

### Optional Environment Variables

Use environment variables when you want the MCP server to forward your Delora API key or use a custom upstream API URL.

| Variable         | Description                                               |
| ---------------- | --------------------------------------------------------- |
| `DELORA_API_KEY` | Optional Delora API key forwarded upstream as `x-api-key` |
| `DELORA_API_URL` | Optional override for the Delora API base URL             |
| `MCP_TRANSPORT`  | `stdio` (default) or `http`                               |
| `PORT`           | HTTP port when `MCP_TRANSPORT=http`                       |
| `HOST`           | HTTP bind address when `MCP_TRANSPORT=http`               |

Example:

```bash theme={null}
# macOS / Linux
DELORA_API_KEY=YOUR_API_KEY npx -y @deloraprotocol/mcp@latest
```

```powershell theme={null}
# Windows PowerShell
$env:DELORA_API_KEY="YOUR_API_KEY"
npx -y @deloraprotocol/mcp@latest
```

<Info>
  `DELORA_API_KEY` is the correct way to provide a key in `stdio` mode. In HTTP mode, you can also send `x-api-key` or `Authorization: Bearer ...` as incoming request headers when your MCP client supports custom headers.
</Info>

### Installation from Source

Clone and build manually:

```bash theme={null}
git clone https://github.com/DeloraProtocol/delora-mcp.git
cd delora-mcp
npm install
npm run build
```

### Running the Server

stdio mode (default):

```bash theme={null}
npx -y @deloraprotocol/mcp@latest
# or from a cloned repository
npm start
```

HTTP mode:

```bash theme={null}
MCP_TRANSPORT=http npx -y @deloraprotocol/mcp@latest
# or from a cloned repository
npm run start:http

# Custom Port
MCP_TRANSPORT=http PORT=3001 npx -y @deloraprotocol/mcp@latest
```

### Configuration for Agent Frameworks

stdio configuration (local agents):

```json theme={null}
{
  "delora": {
    "type": "stdio",
    "command": "npx",
    "args": ["-y", "@deloraprotocol/mcp@latest"],
    "env": {
      "DELORA_API_KEY": "YOUR_API_KEY"
    }
  }
}
```

<Info>
  If your MCP client supports per-server environment variables, set `DELORA_API_KEY` there. This keeps the key on the MCP server side and works for both local `stdio` usage and self-hosted MCP servers.
</Info>

Streamable HTTP:

```json theme={null}
"delora": {
  "type": "streamable-http",
  "url": "http://localhost:3000/mcp",
  "headers": {
    "x-api-key": "YOUR_API_KEY"
  }
}
```

You can also use:

```json theme={null}
"headers": {
  "Authorization": "Bearer YOUR_API_KEY"
}
```

if your MCP client supports custom HTTP headers.

#### Claude Code (CLI & IDE plugins)

Add the MCP server:

```bash theme={null}
claude mcp add delora npx -- -y @deloraprotocol/mcp@latest
# OR
claude mcp add delora --transport http http://127.0.0.1:3000/mcp
```

Verify the connection:

```bash theme={null}
claude mcp list
```

#### Cursor

Add to `.cursor/mcp.json` (project-level) or `~/.cursor/mcp.json` (global):

```json theme={null}
{
  "mcpServers": {
    "delora": {
      "command": "npx",
      "args": ["-y", "@deloraprotocol/mcp@latest"]
    }
  }
}
```

#### GitHub Copilot (VS Code Chat)

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

```json theme={null}
{
  "mcpServers": {
    "delora": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@deloraprotocol/mcp@latest"]
    }
  }
}
```

#### Windsurf

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

```json theme={null}
{
  "mcpServers": {
    "delora": {
      "command": "npx",
      "args": ["-y", "@deloraprotocol/mcp@latest"]
    }
  }
}
```

### Development

```bash theme={null}
# Run in dev mode (no build needed)
npm run dev

# Run local HTTP mode in dev
npm run dev:http

# Build production output
npm run build
```

## 2. Use the Delora Hosted MCP Server

The fastest way to get started is by connecting to the hosted Delora MCP endpoint. No local installation required.

MCP endpoint:

```
https://mcp.delora.build/mcp
```

<Info>
  The hosted endpoint works without local setup. If your MCP client supports custom HTTP headers, you can pass `x-api-key` or `Authorization: Bearer ...` directly to `https://mcp.delora.build/mcp`. If your client does not support headers, use a local or self-hosted Delora MCP server with `DELORA_API_KEY`.
</Info>

Below are setup instructions for supported AI tools.

### Claude Desktop

Add to your Claude Desktop configuration file:

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

```json theme={null}
{
  "mcpServers": {
    "delora": {
      "type": "http",
      "url": "https://mcp.delora.build/mcp"
    }
  }
}
```

### Claude Code

Add via CLI:

```bash theme={null}
claude mcp add delora --transport http --url https://mcp.delora.build/mcp
```

Or create a `.mcp.json` file in your project root:

```json theme={null}
{
  "servers": {
    "delora": {
      "type": "http",
      "url": "https://mcp.delora.build/mcp"
    }
  }
}
```

### Cursor

Add to `.cursor/mcp.json` (project-level) or `~/.cursor/mcp.json` (global):

```json theme={null}
{
  "mcpServers": {
    "delora": {
      "url": "https://mcp.delora.build/mcp"
    }
  }
}
```

### Testing Your Setup

You can verify that the MCP server is running correctly using the [MCP Inspector](https://github.com/modelcontextprotocol/inspector):

```bash theme={null}
npx @modelcontextprotocol/inspector --url https://mcp.delora.build/mcp
```

This will open a web UI at `http://localhost:6274` where you can browse and interact with all available tools.
Alternatively, you can test your AI tool by sending a simple request to the server. If it is configured correctly, it will respond with the expected tool outputs.