FlowGenX

Interactive CLI to configure FlowGenX Productivity MCP across AI coding tools and automatically install Worker AI skills.

One command to connect your AI assistant to 100+ app integrations.

Quick Start

npx flowgenx

That's it. The CLI will:

1. Detect installed AI tools on your system
2. Configure MCP connection for each selected tool
3. Install Worker AI skills automatically

Prerequisites

1. Create a FlowGenX Account

Go to flowgenx.ai and sign up.

2. Access Worker AI

After signing in, navigate to Worker AI from the dashboard. Worker AI is where you connect your apps and get your MCP credentials.

3. Connect Your Apps

In Worker AI, connect the apps you want to use (Slack, HubSpot, Gmail, GitHub, etc.). Each connected app exposes API endpoints that become available through MCP.

4. Get Your MCP URL & API Key

In Worker AI, go to Settings > MCP Configuration to find:

• MCP URL — Your unique endpoint, e.g. https://mycompany-default.work.flowgenx.ai/mcp
• API Key — Your authentication key, e.g. mycompany-consumer-abc123

5. Run the CLI

# npm
npx flowgenx

# bun
bunx flowgenx

# or install globally
npm i -g flowgenx
flowgenx

What Happens When You Run It

┌  FlowGenX MCP Setup
│
◆  Configuration scope → Project or Global
│
◇  Detecting installed AI tools...
│  Found 3 tools: Claude Code, Codex, Gemini CLI
│
◆  Select tools to configure
│
◇  FlowGenX MCP URL → https://your-tenant.work.flowgenx.ai/mcp
│
◇  API Key → your-api-key
│
✓  Claude Code → .mcp.json
✓  Codex → ~/.codex/config.toml
✓  Gemini CLI → ~/.gemini/settings.json
│
✓  Worker AI skills installed
│
└  Setup complete!

Supported AI Tools

| Tool | Config File | Transport | |------|------------|-----------| | Claude Code | .mcp.json / ~/.claude/settings.json | HTTP | | Claude Desktop | claude_desktop_config.json | Stdio (via bundled proxy) | | Codex (OpenAI) | ~/.codex/config.toml | HTTP | | Gemini CLI | ~/.gemini/settings.json | HTTP | | OpenCode | ~/.config/opencode/opencode.json | HTTP | | KiloCode | ~/.config/kilo/kilo.jsonc | HTTP |

Works on macOS, Linux, and Windows.

Skills (Auto-Installed)

The CLI automatically installs FlowGenX Worker AI skills which teach your AI assistant how to use the MCP tools effectively:

| Skill | What it does | |-------|-------------| | flowgenx-help | Getting started, tool overview, auth info | | flowgenx-apps | Discover connected apps and endpoints | | flowgenx-search | Natural language endpoint search | | flowgenx-schema | Inspect endpoint parameters and schemas | | flowgenx-execute | Execute single or batch API calls | | flowgenx-workbench | Python sandbox with tool helpers | | flowgenx-workflow | Full workflow: discover → search → inspect → execute |

Skills are installed globally via npx skills add. To install skills separately:

npx skills add https://github.com/FlowGenX-AI/flowgenx-skills

Available MCP Tools

Once configured, your AI assistant gets access to:

| Tool | Description | |------|-------------| | list_apps | Discover your connected apps | | list_tools | Browse endpoints for a specific app | | search_mcp_tools | Natural language search across all endpoints | | tool_summary | Quick summaries for endpoint IDs | | get_schema | Inspect endpoint parameters and body schema | | execute_mcp_tool | Execute an API call with automatic auth | | multi_execute | Batch execute up to 50 calls in parallel | | workbench_execute | Run Python code in a sandbox with tool helpers | | help | Usage documentation |

Manual Configuration

If you prefer to configure manually instead of using the CLI:

{
  "mcpServers": {
    "flowgenx": {
      "type": "http",
      "url": "https://your-tenant.work.flowgenx.ai/mcp",
      "headers": {
        "apiKey": "your-api-key"
      }
    }
  }
}

[mcp_servers.flowgenx]
url = "https://your-tenant.work.flowgenx.ai/mcp"

[mcp_servers.flowgenx.http_headers]
apiKey = "your-api-key"

{
  "mcpServers": {
    "flowgenx": {
      "httpUrl": "https://your-tenant.work.flowgenx.ai/mcp",
      "headers": {
        "apiKey": "your-api-key"
      }
    }
  }
}

{
  "mcp": {
    "flowgenx": {
      "type": "remote",
      "url": "https://your-tenant.work.flowgenx.ai/mcp",
      "headers": {
        "apiKey": "your-api-key"
      }
    }
  }
}

{
  "mcpServers": {
    "flowgenx": {
      "type": "remote",
      "url": "https://your-tenant.work.flowgenx.ai/mcp",
      "headers": {
        "apiKey": "your-api-key"
      }
    }
  }
}

Author

Mohammad Ismail — [email protected]

License

MIT