I Created an Openai API Server, Because There Wasn’t One

Author(s): Fabio Matricardi

Originally published on Towards AI.

I analyzed OpenVINO’s complexities and built my own API Server in 20 hours — from the idea to the code: the hidden truth about coding with Claude as an AI assistant.

This member-only story is on us. Upgrade to access all of Medium.

Vibe coding, or Coding with AI assistant… or other hundreds names, you call it.

I tested it, because I struggled to install the ovms (OpenVINO Model Server (OVMS)): it was far too complicated. Probably would have worked out-of-the-box with the Docker container, but I really don’t want to use Docker on my Laptop.

After spending 20 hours with the online Documentation and OpenVINO Notebooks, I gave up on Intel ovms, but I gained some motivation to Do It Myself.

Bottom Line Up Front… no tricks no hooks, here in two lines the shocking truth of this Human-AI partnership:

A good Pareto ratio (80% Claude, 20% me) and a more realistic effort ratio (80% me, 20% Claude 3.7).

In this article, I will describe the process (and the results) to write a Python code with Claude Sonnet 3.7 as an assistant.

my personal attempt to openai API compatible openvino server – fabiomatricardi/openvino-api-server

github.com

It will be a 3-episode series… but in Asimov style, every article can be read standalone. The saga sounds like this

how to go from the idea to the codehow to improve the code and include new featureshow to publish it — a pip package ???

Disclaimer:… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI