This is an example of how to use chatbots programmatically without needing a token or api key to get a response from the LLM. This is only restricted to the owner of the chatbot, if the owner shutdown the project we won't be able to get a reponse.
Hugging face can host data science projects that uses models and Gradio as interface, this is hosted in something known as 'space' in hugging face, a project owner can turn on or off his project and decide to make it public or private, huggingface have a free quota for computing usage.
Gradio applications provide an integrated api, to control the app programmatically, it can be used in Python or JS Gradio Client Docs