Neat is a culmination of my research in systems architectures in the last few months. It stands for Network Environment Architecture Tools.

It fuses Runtime + OpenTelemetry to construct a deterministic model for agents to query, so they're more accurate, cheaper & act on the truth.

My thesis is that software tooling & hosting has not evolved even nearly at the same pace as AI. Neat is the answer to that problem. LLMs treat their tokens as the whole truth; our graph gives them the truth, full-stack, by providing a suite of context & control tools to truly reason & act like engineers.

The world of graph context has moved rapidly since I started my research. I was a bit of a perfectionist, however my cofounders rushed me to begin MVP development. This release is the most robust MVP to date. I would like to thank them for everything.

Thank you for reading. Questions & Input are greatly appreciated.

- Deniz, Co-Founder at NEAT.

Working on this opened my eyes to real systems development. AI is still miles behind where it ought to be, and I think we've brung AI closer, even if by a little, to that AI dream future.

Check it out.

- Sinan, Co-Founder at NEAT.