xAI Cookbooks

A collection of pragmatic, real-world examples guiding you from basic to advanced use of xAI's Grok APIs for practical, powerful applications.

How to Run the Notebooks Yourself

Below are steps to run these notebooks locally with your own API key, which you can obtain over at the xAI Console. If you’d rather not run them, all notebook cells are preserved with outputs intact, so you can follow along just by reading.

Prerequisites

• Install uv for managing Python dependencies.

Setup

1. Clone this repo:

   git clone https://github.com/xai-org/xai-cookbook.git
   cd xai-cookbook
   
   

2. Install dependencies:

   uv sync
   
   

3. Install pre-commit hooks:

   uv run pre-commit install
   
   

4. Configure git lfs (install it if you haven't already):

   git lfs install
   
   

5. Copy the .env.example file and rename it to .env and add your API key (the repo is set-up to ignore .env so that it won't be committed).:

   cp .env.example .env
   
   

6. Launch Jupyter in the notebook directory and run the desired notebook:

   uv run jupyter notebook examples/
   

Alternatively, use VS Code’s notebook extension:

• Open the notebook in VS Code.
• Set the Python interpreter to .venv (created by uv sync) via the interpreter selector.
• Run directly in VS Code.

API Key Security

Never hardcode API keys (e.g., XAI_API_KEY=my-key), even during development. This risks accidental commits. Instead, use the .env file. In your notebook, load the key like this:

from dotenv import load_dotenv
import os

load_dotenv()
XAI_API_KEY = os.getenv("XAI_API_KEY")

Git LFS for Large Files

This repo uses Git LFS to track large data files (e.g., images, large JSON) used by some notebooks. To work with these files:

Install Git LFS:

git lfs install

Files are tracked via .gitattributes, check or update it if adding new large files.

Pre-commit Hooks

This repo includes a pre-commit hook with Gitleaks to block accidental API key commits. It’s not foolproof though, if you accidentally commit your key, revoke it immediately in the xAI Console and generate a new one.

There’s also a pre-commit hook to verify that your notebooks run end-to-end without errors.

It’s fine to skip these hooks during development using git commit --no-verify, especially if your notebook makes frequent requests to the xAI APIs, skipping can help preserve your API credits. That said, we strongly encourage you to run your changes with the hooks enabled at least once before submitting a pull request or responding to comments on an open PR. This ensures code quality and catches potential issues. To manually trigger the hooks, use pre-commit run after staging your changes. (e.g. uv run pre-commit run --files examples/your_notebook.ipynb)

Contributing

Want to add a notebook? See CONTRIBUTING.md for guidelines. Key steps:

• Develop your notebook with uv run jupyter notebook examples/.
• Test it runs end-to-end.
• If adding large files (e.g., images, JSON), update .gitattributes for Git LFS.
• Submit a PR!