Build with AI

In this guide, you will learn how to leverage AI coding assistants to help you set up and integrate Kapa.

Kapa's MCP

Kapa provides a hosted MCP server that gives AI assistants access to Kapa's documentation, so they can help you implement Kapa correctly.

This is particularly useful for implementation tasks where you're coding:

• Widget configuration: Let your editor look up styling options and configure the widget for you.
• Agent SDK integration: Build in-product agents while your editor looks up how to use the Agent SDK.
• Chat SDK integration: Build custom chat interfaces while your editor looks up how to use the Chat SDK.

MCP for your users

Want to offer this for your own users? Kapa can power a hosted MCP server for your knowledge sources. See MCP.

Available tool

The Kapa MCP server exposes the following tools to your AI assistant:

Tool	Description
search_kapa_knowledge_sources	Search Kapa's documentation, API reference, and blogs, and return the most relevant content.

Set up

Before you can use the MCP server, you need to install it in your AI assistant.

URL: https://kapa-docs.mcp.kapa.ai

Setup steps vary depending on which client you're using.

Claude Code

Run the following command in your terminal:

claude mcp add --transport http kapa-ai-docs https://kapa-docs.mcp.kapa.ai

Then run the /mcp command in Claude Code and follow the steps in your browser to authenticate.

For more information, see the Claude Code MCP documentation.

Claude Desktop

Add to your Claude Desktop config file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "kapa": {
      "command": "npx",
      "args": ["mcp-remote", "https://kapa-docs.mcp.kapa.ai"]
    }
  }
}

Restart Claude Desktop for changes to take effect.

For more details, see the Claude Desktop documentation.

ChatGPT Desktop

ChatGPT Desktop supports MCP servers in developer mode:

1. Open ChatGPT Desktop.
2. Go to Settings > Features.
3. Enable Developer mode.
4. Navigate to Settings > MCP Servers.
5. Click Add Server and enter:
   • Name: kapa
   • URL: https://kapa-docs.mcp.kapa.ai

For more information, see the ChatGPT Desktop MCP documentation.

Codex

Install MCP: codex mcp add kapa-docs -- https://kapa-docs.mcp.kapa.ai/.

The ~/.codex/config.toml file can also be manually configured:

~/.codex/config.toml

[mcp_servers.kapa-docs]
command = "https://kapa-docs.mcp.kapa.ai"

Note, as of version 0.112 of Codex contains a bug that affects Internal projects. This will cause authentication to fail. A workaround for this is to take the authorisation link generated by Codex and modify it to add the following parameter: &resource=https%3A%2F%2F<YOUR_MCP>.mcp.kapa.ai%2F

Cursor

Add the following to your .cursor/mcp.json file:

.cursor/mcp.json

{
  "mcpServers": {
    "kapa": {
      "type": "http",
      "url": "https://kapa-docs.mcp.kapa.ai"
    }
  }
}

For more information, see the Cursor MCP documentation.

VS Code

Prerequisites: VS Code 1.102+ with GitHub Copilot enabled.

Create an mcp.json file in your workspace .vscode folder:

.vscode/mcp.json

{
  "servers": {
    "kapa": {
      "type": "http",
      "url": "https://kapa-docs.mcp.kapa.ai"
    }
  }
}

For more details, see the VS Code MCP documentation.

Antigravity

To use the Kapa MCP in your Antigravity, follow the instructions below:

1. Open Antigravity.
2. In the Agent sidebar, press ... and then MCP Servers.
3. Press Manage MCP Servers and then View raw config.
4. Paste in the JSON below and save the file.
5. Click "refresh".
6. When prompted, complete the authentication flow.

mcp-config.json

"mcpServers": {
  "kapa-docs": {
    "command": "npx",
    "args": [
      "-y",
      "mcp-remote",
      "https://kapa-docs.mcp.kapa.ai"
    ]
  }
}

Windsurf

To use the Kapa MCP in your Windsurf, follow the instructions below:

1. In the Agent sidebar, press ... and then Open MCP Config File.
2. Paste in the JSON below and save the file.
3. Reload Windsurf.
4. When prompted, complete the authentication flow.

mcp-config.json

{
  "mcpServers": {
    "kapa": {
      "args": [
        "mcp-remote",
        "https://kapa-docs.mcp.kapa.ai"
      ],
    "command": "npx",
    "disabled": false
    }
  }
}

Zed

Install the mcp-remote package from npm: npm install -g mcp-remote

Run mcp-remote to authenticate with Kapa: npx mcp-remote https://kapa-docs.mcp.kapa.ai. This will prompt you to authenticate the session. Your session will automatically get saved in ~/.mcp-remote/

Modify the ~/.config/zed/settings.json settings file to include a context_servers block:

~/.config/zed/settings.json

{
  "context_servers": {
    "kapa": {
      "source": "custom",
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://kapa-docs.mcp.kapa.ai"
      ],
      "env": {}
    }
  }
}

For more details, see the Zed MCP documentation.

Other

MCP is an open protocol supported by many clients. Use the server URL https://kapa-docs.mcp.kapa.ai and refer to your client's documentation for setup instructions.

Most clients accept the standard MCP JSON configuration format:

{
  "mcpServers": {
    "kapa": {
      "url": "https://kapa-docs.mcp.kapa.ai"
    }
  }
}

Kapa Skills

Kapa provides a collection of AI agent skills that you can install in agents like Claude Code, Cursor, and Codex. Each skill represents a recommended workflow for working with Kapa data and capabilities, allowing you to instruct your AI agents to work more effectively with Kapa.

• Analyze Coverage Gaps: Work through topics where your documentation has no coverage and create the missing content directly in your repository.
• Analyze Top Questions: Work through your most common topics and ensure your documentation covers all the angles in which users are asking about them.

See the kapa-skills repository for installation instructions.