Aviro

Environments for Long Horizon Tool Use

Y CombinatorSpring 2025ActiveB2BReinforcement Learning

What they do

Problem

AI models lack reliable environments to train and evaluate long-horizon tool use for high-stakes, real-world tasks across web and enterprise workflows.

Solution

Aviro provides reinforcement-learning environments for training and testing long-horizon tool-using agents across ML research, live web interaction, and enterprise knowledge work.

Who it's for

Frontier AI labs and large enterprises (e.g., Fortune 100) training and evaluating models as tool-using operators.