Cloudflare AI Python Package

This package provides a simple and comprehensive way to interact with Cloudflare Workers AI from Python.

Installation

Install the package using pip:

pip install cloudflare-ai-python

Or from a local directory:

pip install -e .

Usage

Initialize the Client

from cloudflare_ai_python import CloudflareAI

# Replace with your actual values
account_id = "your_account_id"
api_token = "your_api_token"

client = CloudflareAI(account_id, api_token)

List Available Models

models = client.list_models()
for model in models['result']:
    print(f"Model: {model['name']}, Description: {model.get('description', 'N/A')}")

Run a Model

response = client.run_model("@cf/meta/llama-3.1-8b-instruct", "Where did the phrase Hello World come from")
print(response['result']['response'])

Backward Compatibility

For backward compatibility, you can still use the old function:

from cloudflare_ai_python import call_cloudflare_ai

response = call_cloudflare_ai(account_id, api_token, "@cf/meta/llama-3.1-8b-instruct", "Hello World")

API Reference

CloudflareAI Class

• __init__(account_id, api_token): Initialize the client.
• list_models(): Returns a list of available models with detailed information.
• run_model(model, prompt, **kwargs): Run a specific model with a prompt and optional parameters.

Parameters

• account_id: Your Cloudflare Account ID
• api_token: Your Cloudflare API Token
• model: The model identifier (e.g., "@cf/meta/llama-3.1-8b-instruct")
• prompt: The text prompt to send to the model
• **kwargs: Additional parameters like max_tokens, temperature, etc.

Getting Your API Credentials

1. Go to the Cloudflare Dashboard
2. Select "Use REST API"
3. Click "Create a Workers AI API Token"
4. Review the prefilled information and select "Create API Token"
5. Copy the API Token and Account ID

Important: Ensure your API token has the following permissions:

• Workers AI - Read
• Workers AI - Edit

If you get a 401 Unauthorized error, your token may be invalid or expired. Create a new one following the steps above.

Development

Setting up PyPI Publishing

To automatically publish to PyPI when pushing to main branch, you need to configure authentication. Choose one of the following methods:

Method 1: API Token (Recommended)

1. Create a PyPI account at https://pypi.org/
2. Go to Account Settings → API tokens → Add API token
3. Create a token with scope "Entire account (all projects)"
4. Copy the token
5. In your GitHub repository, go to Settings → Secrets and variables → Actions
6. Add a new repository secret named PYPI_API_TOKEN with the token value

Method 2: Trusted Publishing

1. First, manually publish your package to PyPI once using Method 1 above
2. Go to your project on PyPI → Publishing
3. Click "Add publisher" → "GitHub"
4. Configure with these settings:
   • Repository name: him6794/cloudflare-ai
   • Workflow name: publish.yml
   • Environment name: Leave empty
5. The workflow will then use Trusted Publishing automatically

Running Tests Locally

pip install -e .[dev]
pytest

Manual Publishing (for testing)

If you want to test publishing manually before setting up automated publishing:

# Install twine
pip install twine

# Build the package
python -m build

# Upload to PyPI (you'll be prompted for credentials)
twine upload dist/*

Or use the provided script:

python publish_manual.py