A starter kit (template + CLI) that generates a micro-SaaS project for an agentic tool, including a landing page (via prompt-to-app), an MCP server for core functionality, extensibility via agentic execution embed, and deployment instructions for solo developers.

Quick Start • Features • Examples • Contributing

What is this?

AgenticStarter is a CLI‑driven starter kit that scaffolds a sellable micro‑Saas around an agentic tool. It creates the project structure, a static landing page, and an MCP server so solo developers can go from prototype to product in minutes.

$ agenticstarter init my-agentic-tool
Created directory my-agentic-tool with pyproject.toml, src/, and tests/

Problem

Solo developers who build agentic tools struggle with the boilerplate of creating a sellable product (landing page, auth, payments, deployment), causing them to abandon monetization efforts despite having useful tools.

Features

Feature	Description
Project Scaffold	Generates a complete Python project layout with pyproject.toml, source package, and test directory via agenticstarter init.
Landing Page Generator	Produces a static HTML landing page with product name, description, and call‑to‑action button using agenticstarter landing-page.
MCP Server	Starts a minimal Model Context Protocol server on localhost:8000 that handles JSON‑RPC 2.0 messages and supports agentic execution embed via agenticstarter mcp-server start.
CLI Entrypoint	Provides a unified command line interface with subcommands init, landing-page, and mcp-server.
Deployment Instructions	Includes a ready‑to‑use markdown guide for deploying the generated micro‑Saas to common platforms.

Quick Start

1. Clone the repository:
   git clone https://github.com/m2ai-portfolio/agenticstarter.git
2. Enter the directory and install the package in editable mode:
   cd agenticstarter && pip install -e .
3. Scaffold a new micro‑Saas project:
   agenticstarter init my-agentic-tool
   This creates my-agentic-tool/ with the required files.
4. Generate a landing page for your product:
   cd my-agentic-tool && agenticstarter landing-page --name "My Tool" --desc "Awesome tool"
   Produces index.html containing the heading, description, and a Get Started button.
5. Launch the MCP server to test core functionality:
   agenticstarter mcp-server start
   The server listens on port 8000 and replies with a pong message to a ping request.

Examples

Initialize a new project

$ agenticstarter init summary-bot
Created directory summary-bot
 summary-bot/
   pyproject.toml
   src/
   tests/

Create a landing page with custom text

$ agenticstarter landing-page --name "Summary Bot" --desc "Turn long articles into brief bullets"

Resulting index.html (excerpt):

<h1>Summary Bot</h1>
<p>Turn long articles into brief bullets</p>
<a href="#" class="cta-button">Get Started</a>

Run the MCP server and send a ping
In one terminal:

$ agenticstarter mcp-server start
MCP server listening on ws://localhost:8000

In another terminal (using a simple WS client):

$ wscat -c ws://localhost:8000
> {"jsonrpc":"2.0","method":"ping","id":1}
< {"jsonrpc":"2.0","result":"pong","id":1}

File Structure

AgenticStarter: Micro-SaaS Kit for Agentic Tools/
  agenticstarter/          # Core package
    cli.py                 # Click‑based command line interface
    project_template.py    # Scaffold generation logic
    landing_page_template.py # HTML landing page creator
    mcp_server.py          # Minimal MCP server implementation
    deployment_instructions.md # Deployment guide for solo devs
  tests/                   # Unit test suite
    test_project_scaffold.py
    test_landing_page.py
    test_mcp_server.py
  assets/                  # Visual assets
    infographic.png
  screenshots/             # Example outputs and verification screens
  agenticstarter.py        # Console script entry point
  pyproject.toml           # Project metadata and dependencies
  README.md
  LICENSE
  .gitignore

Tech Stack

Technology	Purpose
Python 3.11+	Core language runtime
click	Building the CLI interface
pytest	Running unit tests

Contributing

Fork the repo, make your changes, run pytest to verify, then submit a pull request.

License

MIT

Author

Matthew Snow -- M2AI | @m2ai-portfolio