Wordware

The easiest way to get started is using the interactive installation process with npx:
- Entering your Wordware API key
- Setting up Claude configuration (optional)

- Prompt you for configuration details if not provided
- Create necessary configuration files
- Set up your local environment to run Wordware apps

After running the npx command, you can start the MCP server with:

If you prefer to install the package permanently:

# Install globally from npm registry npm install -g wordware-mcp # Or install locally in your project npm install wordware-mcp # Or clone this repository and install globally git clone https://github.com/{username}/wordware-mcp.git cd wordware-mcp npm run install-global

Using npx directly (no installation required)

You can run wordware-mcp using npx without installing it first:

# Interactive mode - will prompt for required information npx wordware-mcp # Or with command line parameters npx wordware-mcp --api-key your-api-key --port 3000 # Start MCP server after configuration npx wordware-mcp start

If installed globally, you can run in one of two ways:

# Option 1: Create an .env file in your current directory first (see Configuration section) wordware-mcp # Option 2: Pass parameters directly via command line wordware-mcp --api-key your-api-key --port 3000

Options: --api-key, -k <key> Wordware API key (required unless in .env file) --port, -p <port> Port to run the server on (default: 3000) --help, -h Show this help message

// In your script import { startMCP } from "wordware-mcp"; // Start the MCP server startMCP();

You can configure the MCP server in two ways:

Create a.envfile with the following variables:

Pass the configuration directly when running the command:

For your wordware flows to be accessible via MCP, you'll need to create an API key. For that, click on your profile picture in the top right corner > API keys > Create a new key > Copy your key

For your app to be triggered as MCP, you'll need to deploy it. To do that, head to your app. You should see a "Deploy" button in the top right corner. Then head to the deployment page.

On the deployment page, you'll see your deployment url:https://app.wordware.ai/explore/apps/{app_id}. Get your app_id from there

To use this MCP server with Claude Desktop:
-

Make sure Claude for Desktop is installed

Modify the Claude desktop config file located at:~/Library/Application\ Support/Claude/claude_desktop_config.json

{ "mcpServers": { "wordware": { "command": "wordware-mcp" } } }

Here's a complete workflow example to get up and running quickly:

# Run the interactive setup npx wordware-mcp # Follow the prompts to: # - Enter your Wordware API key # - Configure Claude integration (if desired) # Once configured, start the server npx wordware-mcp start

After starting the MCP server, your Wordware apps will be accessible at:

You can trigger your Wordware flows via HTTP requests:

// Example: Calling your Wordware app from JavaScript async function callWordwareApp() { const response = await fetch("http://localhost:3000/api/run/your-app-id", { method: "POST", headers: { "Content-Type": "application/json", }, body: JSON.stringify({ // Your input data here prompt: "Your prompt to the AI model", // Any other parameters your Wordware app expects }), }); const result = await response.json(); console.log(result); }

During development, any changes you make to your Wordware apps will be immediately available - just refresh your app or make a new API call.

If you want to contribute to this package:

# Clone the repository git clone https://github.com/yuhuangou/wordware-mcp.git cd wordware-mcp # Install dependencies npm install # Build the package npm run build # Run in development mode npm run dev

-

"Command not found" after installation

If you seecommand not foundafter installing with npx:

# Make sure the package is installed globally npm install -g wordware-mcp # Check your npm global path is in your PATH npm config get prefix # Add the resulting path + /bin to your PATH if needed

If your configuration isn't being detected:

# Check if .env file exists in current directory ls -la .env # Manually run with parameters to bypass .env npx wordware-mcp --api-key your-api-key

If you see connection errors when trying to use your apps:

# Check if server is running lsof -i :3000 # Restart server with verbose logging npx wordware-mcp start --verbose

If you encounter permissions errors with npx:

# Run with sudo (not recommended as permanent solution) sudo npx wordware-mcp # Fix npm permissions chown -R $(whoami) $(npm config get prefix)/{lib/node_modules,bin,share}

For more assistance, please file an issue on our GitHub repository.

The following environment variables can be set in the.envfile:

- PORT- The port to run the server on (default: 3000)
- WORDWARE_API_KEY- Your Wordware API key

This is a web browser that enables your coding agent, such as Claude Code, to visit websites on your behalf and assist you in identifying bugs or creating UI test cases.

What Shopify did for ecommerce, Chipp does for AI agents. Build, deploy, and monetize AI agents for your business — no engineering team required.

CodeVF MCP lets AI hand off problems to real engineers instantly, so your workflows don’t stall when models hit their limits.

Client implementation for Mastra, providing seamless integration with MCP-compatible AI models and tools.

Agent-native developer Q&A API with MCP + A2A endpoints for citations, job pickup, and answer submission.

On-demand access to 150+ specialist AI agent templates — search, browse, and spawn agents. 150x reduction in context usage vs loading agents locally.

AgentChatBus is a persistent AI communication bus that lets multiple independent AI Agents chat, collaborate, and delegate tasks — across terminals, across IDEs, and across frameworks.

Remote MCP server (Streamable HTTP) at https://mcp.agenticrail.nz/ — deterministic step-order enforcement for AI agents. evaluate_step returns ALLOW or DENY before a step runs; verify_receipt proves a sequence's Ed25519-signed, hash-chained receipt chain is intact. No auth required: omit the bearer token and calls run on the public demo key. That first clause matters — the form has no "remote/hosted" field, and putting the endpoint in the description is the convention on that list ("Fully REMOTE! Just use…"). The rest mirrors your own server card verbatim, so the listing and the card can't drift.

An AI Agent with optional Human-in-the-Loop Safety and Model Context Protocol (MCP) integration.

Agent Mail Rooms is a paid remote MCP endpoint for OpenAI Codex agent coordination MCP. It exposes structured JSON tools, a public server card, token-based access, usage receipts, and audit-