Lotus MCP

Name: Lotus MCP
Author: chrisvin-jabamani

Create browser automation skills by demonstration. Works with Cursor and Claude Code.

What is Lotus?

Lotus lets you create reusable browser automations (called "skills") just by showing what you want to do. Instead of writing code, you demonstrate the workflow in a browser while narrating what you're doing, and Lotus converts it into an executable skill.

Quick Start

For Cursor Users

1. Install the MCP:

npm install -g github:chrisvin-jabamani/lotus-mcp

2. Install the VS Code Extension:

# Download the extension curl -L -o lotus.vsix https://github.com/chrisvin-jabamani/lotus-mcp/raw/main/extension/lotus-0.1.0.vsix # Install in Cursor cursor --install-extension lotus.vsix

3. Get your API key: Sign up at uselotus.ai to get your API key.

4. Configure Cursor MCP settings:

Press Cmd+Shift+J (Mac) or Ctrl+Shift+J (Windows/Linux), go to MCP settings, and add:

{ "mcpServers": { "lotus": { "command": "lotus-mcp-cursor", "env": { "LOTUS_API_KEY": "your-api-key-here", "LOTUS_URL": "https://www.uselotus.ai" } } } }

5. Restart Cursor

6. Start creating skills!

Just ask Claude:

"Create a skill to check court availability on Rec.us"

The browser will open automatically in your Cursor IDE. Demonstrate the workflow, then say "I'm done."

For Claude Code Users

1. Install the MCP:

npm install -g github:chrisvin-jabamani/lotus-mcp

2. Get your API key: Sign up at uselotus.ai to get your API key.

3. Add to Claude Code:

claude mcp add lotus -- lotus-mcp-cli

When prompted, set the environment variables:

LOTUS_API_KEY: Your API key from uselotus.ai
LOTUS_URL: https://www.uselotus.ai

4. Start creating skills!

Just ask Claude:

"Create a skill to check court availability on Rec.us"

The browser will open in Chrome/Safari. Demonstrate the workflow, then come back and say "I'm done."

How It Works

You: "Create a skill to book a tennis court" ↓ Claude calls lotus_start_recording ↓ 🌐 Cloud browser opens (in Cursor panel or system browser) ↓ You: Navigate, click, fill forms while narrating ↓ You: "I'm done" ↓ Claude calls lotus_stop_recording → lotus_refine_skill ↓ ✅ Skill created! Now you can say "Book a tennis court for Saturday"

Available Tools

Tool	Description
`lotus_start_recording`	Start recording a new skill workflow
`lotus_stop_recording`	Stop recording and analyze the workflow
`lotus_refine_skill`	Convert recording into executable skill
`lotus_refine_status`	Check refinement progress
`lotus_refine_answer`	Answer AI clarifying questions
`lotus_cancel_session`	Cancel current session
`lotus_<skill_name>`	Execute a created skill

Troubleshooting

Cursor: Browser doesn't open automatically

Make sure the Lotus extension is installed: Cmd+Shift+P → type "Lotus" → you should see "Lotus: Open Browser"
Try running the command manually: Cmd+Shift+P → "Lotus: Open Browser"
Check that the MCP is configured correctly in settings

Claude Code: Browser doesn't open

Make sure you have a default browser set on your system
Try opening manually - the URL will be printed in the response

"Failed to connect to Lotus"

Check your API key is correct
Check LOTUS_URL is set to https://www.uselotus.ai
Make sure you have internet connectivity

Development

# Clone the repo git clone https://github.com/chrisvin-jabamani/lotus-mcp cd lotus-mcp # Install dependencies npm install # Build npm run build # Test locally LOTUS_API_KEY="your-key" LOTUS_URL="http://localhost:3000" LOTUS_PLATFORM="cli" node dist/index.js