# GenerableLLM

> GenerableLLM lets you run remote AI models from OpenAI, Anthropic, and Cloudflare using Apple's FoundationModels SDK. Your Swift app connects through a secure Cloudflare backend without exposing API keys. Switch models, control costs, and keep using the familiar FoundationModels APIs.

- Kind: Swift package
- Tags: Cloudflare, Claude, Codex
- Price: $10 USD, plus tax where it applies; one payment, no account
- Buy: https://tools.natashatherobot.com/buy/generablellm
- Page: https://tools.natashatherobot.com/generablellm
- Seller: NatashaTheRobot

## What happens after you buy

1. **Download.** Click the download link to get the ZIP file. Extract it, then share the folder with your agent.
2. **Set up with your agent.** Your agent will install the packages, set up the Cloudflare Worker, and check the build. It will then guide you through deploying it to your Cloudflare account and testing the setup with the StoryBook demo app.
3. **Make it yours.** Ask your agent to add AI features, choose models, adjust usage limits, or customize the package and backend to fit your app.

GenerableLLM connects your Swift apps to remote AI models from OpenAI, Anthropic, and Cloudflare through Apple's FoundationModels SDK. Keep using familiar FoundationModels APIs without exposing API keys or building your own backend from scratch. Switch models without updating your app, manage usage and costs, and get started faster with coding agent-assisted setup and a working demo app.

## Features

- **Keep API keys secure.** Access remote AI models without exposing provider credentials in your app.
- **Keep using FoundationModels.** Use remote models with familiar APIs, including streaming, structured responses, and tool calling where supported.
- **Switch models without updating your app.** Choose from different AI providers and configure fallback models.
- **Control usage and costs.** Set spending limits and quotas to manage how your app uses AI.
- **Use Apple's on-device AI first.** Fall back to remote models when Apple's model isn't available.
- **Build and test offline.** Develop and test your AI features without an internet connection or paid model calls.

## FAQs

### Do I need an AI coding agent to use GenerableLLM?

No. GenerableLLM works as a regular Swift package. It also includes instructions for Claude Code, Codex, and other coding agents to help you set it up and deploy the backend.

### Do I need API keys from AI providers?

No. Cloudflare AI Gateway handles access to the models, so you don't need separate API keys from OpenAI, Anthropic, or other supported providers. Usage is billed through your Cloudflare account.

### Which AI models and FoundationModels features are supported?

You can use models supported by Cloudflare AI Gateway, including models from OpenAI, Anthropic, and Cloudflare. GenerableLLM supports streaming and structured responses through FoundationModels. On-device tool calling is currently supported with Gemma 4. The package includes full source code, so you can easily customize it to support additional models and features as your needs change.

## What you get

- `GenerableLLM` and `GenerableWorker` Swift packages
- StoryBook, a demo app with its own Worker

## What you need

- macOS 27 or later
- Xcode 27 or later
- A Cloudflare account

## It is a starter tool

GenerableLLM is a starter tool you can customize for your apps.

- You can use, copy, and change it for yourself and your work, and build it into any number of apps. Your AI agents can read, build, and change it for you.
- The license is for one person. Each person who changes or builds the tool needs their own copy.
- You can share, publish, or sell an app with the tool built in, and deploy its Worker for your app. Your app must do more than GenerableLLM alone.

The full terms are in the `LICENSE` file included with your download. The tool comes without warranty.