REST APIs in the age of AI.

REST → AREST.

Future of APIs → APAIs specification.

No membership? use this link to read this for free! :)

Introduction

The AREST (Adaptive Representational State Transfer) architecture proposes a new paradigm for REST APIs by integrating Artificial Intelligence (AI) and Large Language Models (LLMs) into their foundational operation. Unlike traditional REST APIs that rely on pre-defined databases or fixed processing logic, AREST APIs dynamically generate responses for unknown inputs and enhance their database over time, creating an adaptive and self-evolving and self-healing API system.

This kind of mechanisms have surely been used in clever architectures in the past, but this is an attempt to standardize the usage and propose a new paradigm that can be adopted in a more widely manner as an step toward the future of software in the era of Artificial Intelligence. ∞

Overview

AREST combines traditional REST principles with AI/LLM-driven inference mechanisms. Key features include:

• Dynamic Response Generation: When a query is not found in the database, the API dynamically fetches the response from an AI/LLM service.