About
A practical intelligence layer for distributors.
Built to solve a specific operational problem, not to be the most sophisticated piece of software in the supply chain stack.
Why this exists
The decision problem in distribution.
Ethan Byrne · Founder & CEO
Allodial Supply Co started with a straightforward goal: supply products to local businesses efficiently. The intention was to understand distribution from the inside, covering sourcing, pricing, logistics, and customer behavior, before building software around it.
Through that work, a deeper problem became visible. The businesses being supplied were operating reactively. Not because they couldn't get product, but because no one had built a system to predict when they would need it. Reorders happened when a customer called, when someone noticed the shelf was low, or when an emergency order was already unavoidable.
The cost of this pattern isn't just freight premiums on emergency orders. It's strained customer relationships, capital tied up in dead inventory offsetting the stockouts, and the accumulated friction of running an operation on delayed information. Allodial Predict is the operational intelligence layer built to close that gap.
What guides this
Company
Allodial Predict uses statistical forecasting models running entirely in PostgreSQL, so your inventory predictions never involve external AI. In-app assistance (Allobot) and import column mapping process operational data through OpenAI under its API terms, never to train models. The correct framing is "statistical forecasting" and "advanced forecasting", not "no AI." AI-assisted narrative summaries are generated from aggregated outputs, and your raw inventory records are never used to train any external AI.
Want to run a Replay on your data?
We work directly with distribution operators. Reach out and we'll show you what the model finds in your history.