Ollaya wants to be Ollama for decision models

A new open project called Ollaya is positioning itself as a way to run Jev-style decision models locally. The basic idea is familiar to anyone watching local AI tools. Instead of sending every request to a hosted service, a business could potentially run the relevant decision model inside its own environment. Ollaya is early, but it reached the Hacker News front page quickly, collecting 76 points and 27 comments within hours of launch. You can see how the project describes itself at Ollaya’s site.

Why should an operator care? Decision models point to a different use of AI than drafting emails or summarising documents. The focus is on selecting an action, route, or outcome based on a set of inputs. Think of internal processes where someone currently checks a few conditions, applies a policy, then decides what happens next. Local hosting may matter where data control, reliability, or internal system access are important, though Ollaya’s real-world capability is still something to watch rather than assume.

The broader signal is that tools for routing and running decision models are starting to appear alongside the better-known language model tools. That could make smaller, focused AI systems easier to deploy for operational decisions, not just content tasks. This is the kind of thing we build into an AI command centre, where the value comes from connecting models to a clear business workflow and keeping the decision trail visible.