Skip to main content

Models and APIs

What is an API?​

An API (Application Programming Interface) is the bridge through which UNIChat communicates with artificial intelligence models.

UNIChat itself is a Client, functioning somewhat like an exquisite browser or music player. In order for you to converse with an AI, you need to connect to a Server, which in this text refers to the API.

Imagine:

  • UNIChat is your mobile phone.
  • The API is the telecom operator's cell tower.

Only after inserting a SIM card (configuring an API Key) can the phone make calls (converse with AI).

Why Configure APIs?​

  1. Flexibility: You have the freedom to choose whether to use the world's smartest models (like GPT-4, Claude 3.5) or the most cost-effective ones (like DeepSeek, Gemini).
  2. Privacy and Control: By supporting access from local or custom endpoints, you have full control over where your data flows. You can even run local models (like Ollama) entirely offline.
  3. Cost Efficiency: Accessing model provider APIs directly is generally much cheaper than subscribing to various wrapper services, and it's strictly pay-as-you-go.

In the upcoming sections, we will guide you on how to obtain and configure these APIs.