Welcome

Welcome to AIVAX. Our service makes it easier to develop intelligent AI models that use a knowledge base provided by you to converse with the user, answer questions, provide real-time information, and more.

To get started, all endpoints must be made to the AIVAX production URL:

https://inference.aivax.net/

Concepts and definitions

Understand the concepts used by the API below:

• Account: represents a user account, which has an authentication token.
  • Collection: represents a collection of knowledge documents. A user can have multiple document collections.
    • Document: represents a fact, a single piece of knowledge, and an item in a collection. A collection can have multiple documents.
  • AI Gateway: represents an AI gateway that benefits from or does not use a knowledge collection, such as a plug-and-play knowledge middleware for a model.
    • Embedded model: represents an AI model that AIVAX provides to the user.
    • Chat client: represents a user interface that makes the AI gateway available through an interactive online chat.
      • Chat session: hosts a conversation and context of a chat client.

Handling errors

All API errors return an HTTP response with a non-OK status (never 2xx or 3xx), and always follow the JSON format:

{
    "error": "An explanatory error message",
    "details": {} // an object containing relevant error information. Most of the time it is null
}