MCP Server

Blockend includes a built-in Model Context Protocol (MCP) server that allows AI coding assistants to interact with the Blockend CLI directly.

Instead of manually searching for blocks or running CLI commands yourself, your AI assistant can:

• Discover available Blockend blocks
• Analyze your current project
• Install production-ready backend blocks
• Automate backend scaffolding without leaving your editor

How it works

The MCP server exposes a small set of tools over the standard input/output (Stdio) transport.

When connected, your AI assistant can call these tools instead of asking you to run CLI commands manually.

AI Client
      │
      ▼
Blockend MCP Server
      │
      ▼
Blockend CLI
      │
      ▼
Your Project

Supported AI Clients

Blockend can automatically generate configuration files for the following clients:

Configure MCP

Initialize MCP for your preferred client.

Interactive setup

npx blockend-cli mcp init

You'll be prompted to choose your AI client.

Configure a specific client

npx blockend-cli mcp init --client cursor

Supported values:

claude
cursor
vscode
codex
windsurf

Skip confirmation prompts

npx blockend-cli mcp init --client cursor --yes

Overwrite existing configuration

npx blockend-cli mcp init --client cursor --force

Preview without writing files

npx blockend-cli mcp init --dry-run

This prints the generated configuration without modifying your project.

Starting the MCP server

Once configured, your AI client starts the server automatically.

The command executed is:

npx -y blockend-cli mcp

The server communicates using the standard MCP Stdio transport.

Available Tools

list_blocks

Returns every Blockend block available for installation.

Example prompts:

• "Show all available authentication blocks."
• "List all logging-related blocks."
• "Find caching blocks."

detect_project

Analyzes your current project and detects its runtime environment.

The AI can understand things like:

• Framework
• Runtime
• Package manager
• Project structure

Example prompts:

• "Analyze this project."
• "What framework am I using?"
• "Inspect the current repository."

add_block

Installs a Blockend block directly into your project.

Example prompts:

• "Install the logger block."
• "Add the rate-limit block."
• "Install the error-handler block."

The server internally executes the equivalent of:

npx blockend-cli add <block-name>

Example Workflow

After connecting Blockend to your AI assistant, you can use natural language.

Analyze this project.

↓

The assistant calls detect_project.

What authentication blocks are available?

↓

The assistant calls list_blocks.

Install the logger block.

↓

The assistant calls add_block.

Everything happens directly inside your project without manually running CLI commands.

Generated Configuration

Depending on your selected client, Blockend generates the appropriate configuration automatically.

Example JSON configuration:

{
  "mcpServers": {
    "blockend": {
      "command": "npx",
      "args": ["-y", "blockend-cli", "mcp"]
    }
  }
}

Example TOML configuration (Codex CLI):

[mcp_servers.blockend]
command = "npx"
args = ["-y", "blockend-cli", "mcp"]

CLI Reference

Initialize MCP

npx blockend-cli mcp init

Initialize for a specific client

npx blockend-cli mcp init --client cursor

Preview configuration

npx blockend-cli mcp init --dry-run

Force overwrite

npx blockend-cli mcp init --force

Skip prompts

npx blockend-cli mcp init --yes

Next Steps

After configuring MCP:

1. Restart your AI editor or client if necessary.
2. Open your project.
3. Ask your AI assistant to analyze the project or install Blockend blocks.
4. The assistant will automatically communicate with the Blockend MCP server to perform the requested actions.