Claude on Windows: The “Hello World” That Gets You Building Again (VS Code + Python Guide)

I need to start with something slightly uncomfortable.

I lead major technology programs. AI Gateway. MCP Gateway. Enterprise API platforms. I sit in rooms where we talk about scale, architecture, governance, and multi-million dollar infrastructure decisions.

And yet, it has been a long time since I wrote what I would call meaningful code.

Yes, I run SixSigmaStudyGuide.com. Yes, I occasionally modify PHP snippets or run MySQL queries. But that is WordPress integration work. Plugin orchestration. Configuration.

That is a long way from my roots as a Java N-tier developer.

I have been comfortable using ChatGPT, Copilot, Gemini, Grok. But recent developments around Claude, especially the OpenClaw phenomenon, feel like a step change. The ecosystem is accelerating. Real capability is emerging.

I want first-hand experience. I want to understand it from the inside. And honestly, I want to rediscover the joy of creating things.

None of this would even be possible without the investment I made last year upgrading my super old, clunky desktop. That journey, guided by ChatGPT, is documented here: Lean Thinking Meet ChatGPT.

That hardware upgrade was my infrastructure baseline. This Claude setup is my capability baseline.

The Real Goal: From Claude UI to Local OpenClaw

My immediate goal is modest: get experience with Anthropic Claude through their UI and API.

The long-term goal is more ambitious: eventually progress toward running a local OpenClaw-style instance on my Windows machine.

You do not jump straight to distributed autonomy. You build competence first.

In Six Sigma language, this is the Define and Measure stage of a capability build. If you are newer to structured improvement, review our overview of DMAIC.

Small, controlled experiment. Then scale.

Pre-Work Already Completed

• Created and funded an Anthropic account
• Updated Windows environment variables
• Upgraded pip on my desktop
• Created a GitHub account
• Downloaded and configured VS Code
• Installed Python and required libraries
• Created two “Hello World” scripts
• Debugged errors and executed successfully

None of that is glamorous. All of it is foundational.

Operational excellence is built on disciplined fundamentals.

Step 1: Install Python (Because Nothing Runs Without It)

Everything downstream depends on a clean Python install. If Python is even slightly “off,” you will burn time debugging issues unrelated to Claude.

Download Python (Windows)

Use the official site:

• python.org/downloads/windows
• python.org/downloads

Choose the 64-bit installer for most modern Windows machines.

Install With Correct Settings

• Check Add python.exe to PATH
• Ensure pip is selected
• Enable “Disable path length limit” if prompted

Verify Installation

python --version py --version pip --version

You should see version numbers. If not, restart your terminal or reinstall cleanly.

Upgrade pip

python -m pip install --upgrade pip

Create a Project Folder

mkdir claude-windows-hello cd claude-windows-hello

Step 2: Install VS Code (Clean Execution Environment)

Install from:

• code.visualstudio.com/download

During installation:

• Add to PATH
• Add “Open with Code” to Explorer

Open the Folder (Not Just a File)

cd claude-windows-hello code .

Select Interpreter

Ctrl + Shift + P → Python: Select Interpreter → choose Python 3.12+

Sanity Check

import sys print("Python executable:", sys.executable) print("Python version:", sys.version)

python sanity_check.py

Step 3: Secure the API Key (Governance Matters)

Create and fund your Anthropic account. Generate an API key beginning with:

sk-ant-xxxxxxxxxxxxxxxxxxxxxxxx

Store it in a password manager.

Set Environment Variable (Permanent Recommended)

Windows → Edit Environment Variables → New User Variable:

Name: ANTHROPIC_API_KEY Value: sk-ant-...

Restart VS Code after setting this.

Verify

echo %ANTHROPIC_API_KEY%

If nothing prints, fix it before moving on.

Step 4: Install the Anthropic SDK (Enable Capability)

Create Virtual Environment

python -m venv .venv .\.venv\Scripts\activate

Upgrade pip

python -m pip install --upgrade pip

Install SDK

pip install anthropic

Verify

pip show anthropic

Select .venv Interpreter in VS Code

Ctrl + Shift + P → Python: Select Interpreter → choose the .venv path.

Step 5: Your First Claude API Call

Create hello_claude.py:

from anthropic import Anthropic client = Anthropic() message = client.messages.create( model="claude-sonnet-4-6", max_tokens=200, messages=[ {"role": "user", "content": "Hello Claude! Tell me one fun fact about lobsters."} ], ) print(message.content[0].text)

Run:

python hello_claude.py

If successful, you will see a response printed to the terminal.

Common Early Errors (And How to Diagnose Them Calmly)

404 Model Not Found

anthropic.NotFoundError: 404 model not found

from anthropic import Anthropic client = Anthropic() models = client.models.list() for m in models.data: print(m.id)

Authentication Errors

Re-check:

echo %ANTHROPIC_API_KEY%

Restart VS Code if needed.

Interpreter Mismatch

If you see:

ModuleNotFoundError: No module named 'anthropic'

Select the correct .venv interpreter.

Conclusion: Rebuilding the Builder

This was not really about lobsters.

It was about rebuilding builder muscle.

Upgrade hardware. Install Python cleanly. Configure intentionally. Secure credentials properly. Execute a working API call.

None of it is flashy. All of it matters.

You now have a repeatable, working Claude integration on your local Windows machine.

• You are no longer just consuming AI.
• You are orchestrating it.
• You can integrate it into workflows.
• You can experiment deliberately.

What Comes Next

In the next article, we move from “Hello World” to a practical bookmark classification workflow using Claude.

• Structured input (CSV or JSON)
• Intentional prompt design
• Programmatic response handling
• Token and cost awareness

Hello World proves you can connect. The next project proves you can build.

Capability compounds.