M13 solution

The expected, fully-commented artifact for M13's lab. Peek only after you've tried.

File	What it is
chat_local.py	Chats with an open model running locally via Ollama's local API (POST localhost:11434/api/chat), same shape as M4's hosted call, but no API key, offline, and free. Handles "Ollama not running" gracefully.

There's no .env here on purpose, local models need no key.

Run it

Install Ollama and pull a small model (see the Ollama guide), then with your venv active:

pip install requests
python chat_local.py

How this was verified

Verified on Python 3 with the HTTP call mocked (no running Ollama needed): - compiles; - the client posts to http://localhost:11434/api/chat with the right body (model, messages, stream: false) and correctly parses the reply from message.content, confirmed by inspecting what the mocked requests.post received and returning Ollama's documented response shape; - the "Ollama not running" path (ConnectionError) is handled with a clear message, no crash.

Ollama itself wasn't run here (the build sandbox has no Ollama daemon or downloaded model), > the install + ollama run + live local generation must be piloted on a real machine. The code follows Ollama's documented local API. No API key or billed call was used (local models are free).