RealTask/free-llm-api ? reverse-engineered prompt

Build me a Python tool that makes it easy to use a bunch of free AI APIs from one place.

I want one simple package that can pick the best free provider for a task, like chat, image generation, speech to text, or embeddings, then fall back to another provider if one is down or rate limited. It should support a command line interface and a small web app so I can test providers, see health status, and try example prompts without writing code.

Please include a clean way to store API keys, basic retry and rate limit handling, caching, streaming responses, and some simple examples for chat, image generation, and transcription. If there are current provider details that might have changed, look up the latest docs online and use the best available free options. Also add tests and make it easy to install and run locally.