Prerequisites for Using learn-claude-code: Complete Setup Guide

To use learn-claude-code, you need Python 3.9 or higher, a valid Anthropic API key stored in an .env file, and the dependencies listed in requirements.txt installed via pip.

The shareAI-lab/learn-claude-code repository provides a hands-on framework for building AI agents with Claude. Before you can execute agent loops or launch the optional web interface, you must configure your environment with specific runtime dependencies and API credentials as defined in the source code.

Core System Requirements

Python Runtime and Package Manager

The codebase requires Python 3.9 or newer and a functional pip installation. This is the minimum version supported by the anthropic Python SDK dependency specified in the project.

You also need Git to clone the repository from GitHub.

Anthropic API Credentials

All agent functionality requires external API access. You must obtain a valid Anthropic API key (or compatible provider key) and store it in an environment variable file.

According to the .env.example template in the repository root, create a .env file containing:

• ANTHROPIC_API_KEY – Your secret API key (e.g., sk-ant-...)
• MODEL_ID – The specific model identifier (e.g., claude-sonnet-4-6 or other supported variants)

The python-dotenv>=1.0.0 package (listed in requirements.txt) loads these variables at runtime.

Python Dependencies

Install the required libraries using the lock file provided:

pip install -r requirements.txt

This command installs two critical packages:

• anthropic>=0.25.0 – The official SDK for interacting with Claude models
• python-dotenv>=1.0.0 – For loading environment variables from .env files

Optional Web Interface Prerequisites

Node.js and npm for the Next.js Frontend

If you want to run the visualization web UI located in the web/ directory, you need Node.js 16 or higher and npm. The frontend is built with Next.js and requires separate installation steps from the Python backend.

Reference the web/package.json file for the specific Node dependency tree.

Step-by-Step Installation

Follow these verified steps to satisfy all prerequisites for using learn-claude-code:

1. Clone the repository

   git clone https://github.com/shareAI-lab/learn-claude-code
   cd learn-claude-code

2. Install Python dependencies

   pip install -r requirements.txt

3. Configure environment variables

   cp .env.example .env
   # Edit .env to add your ANTHROPIC_API_KEY and MODEL_ID
   

4. Verify the core agent runs

   python agents/s01_agent_loop.py

   This executes the minimal agent implementation at agents/s01_agent_loop.py, confirming your API credentials and dependencies are functional.

5. Start the web UI (optional)

   cd web
   npm install
   npm run dev

   This launches the Next.js development server on http://localhost:3000 as documented in web/README.md.

Key Configuration Files

The following files define the exact prerequisites and versions required:

• requirements.txt – Specifies anthropic>=0.25.0 and python-dotenv>=1.0.0 as runtime Python dependencies
• .env.example – Template defining required environment variables (ANTHROPIC_API_KEY, MODEL_ID)
• agents/s01_agent_loop.py – Entry point script to verify Python setup and API connectivity
• web/package.json – Lists Node.js dependencies for the optional frontend
• web/README.md – Contains Node.js version requirements and npm commands for the UI

Summary

• Python 3.9+ with pip is mandatory for running any agent code
• Anthropic API key and Model ID must be configured in a .env file copied from .env.example
• pip install -r requirements.txt installs the anthropic and python-dotenv packages
• Git is required only for initial repository cloning
• Node.js 16+ and npm are optional dependencies needed only for the web visualization interface

Frequently Asked Questions

What Python version is strictly required for learn-claude-code?

You need Python 3.9 or higher. The requirements.txt file specifies dependencies compatible with Python 3.9+, and the anthropic SDK requires this minimum version to function correctly with the repository's agent loop implementations.

Is the web interface mandatory to use the framework?

No. The Node.js and npm prerequisites are optional. You can execute all core agent functionality and learning exercises using only the Python runtime and the scripts in the agents/ directory, such as agents/s01_agent_loop.py. The web UI at web/ provides visualization but is not required for the API interactions.

Where do I obtain the Anthropic API key?

You must create an account at the Anthropic developer console (or compatible provider) and generate an API key. Copy this key into the .env file as ANTHROPIC_API_KEY. Without this credential, the agent scripts cannot authenticate with the model endpoints.

Can I use a different model provider instead of Anthropic?

The repository is configured for Anthropic's API by default, as seen in the .env.example file's MODEL_ID variable structure. While the code in agents/s01_agent_loop.py uses the anthropic Python SDK, you could theoretically adapt the client initialization for compatible API endpoints, though this would require modifying the source code directly.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}