Skip to content

Latest commit

 

History

History
30 lines (19 loc) · 1019 Bytes

File metadata and controls

30 lines (19 loc) · 1019 Bytes

Everything API

An AI-powered API that does not exist, yet does everything.

How it works

It works by delegating the response generation to an LLM model running with ollama, somewhere.

Running locally

With the power of Docker, this task becomes simple:

docker compose up 

This will download the start a container with the api, a container with ollama, and another container that's responsible for downloading the desired ollama model. This model is grabbed from the .env file, variable EVERYTHING_API_OLLAMA_MODEL.

What's the point?

This is a simple proof-of-concept of how one can use LLMs easily with Python.

I believe there could be some potential security benefits to make a scraper's life harder by allowing him to call every endpoint imaginable, even if it doesn't actually serve any purpose.

In the realm of API security, obscurity through infinite possibility becomes a feature rather than a flaw.

Schrödinger's API Principle by R1, Deepseek (circa 2025 BC)