Better forecasts. Less stock. Fewer stockouts. Everything we do at OpenForecast — consulting, training, and research — runs on forecasting software we build ourselves and publish as open source. This page explains what those tools do for a business, and where to find them if you want to look under the bonnet.
Why it matters to you
Most demand planning runs on tools that quietly cost money. Standard forecasting engines just extrapolate and break the moment a promotion, price change, or supply disruption enters the picture. Some ERP forecasting modules behave like black boxes, which are hard to understand and explain to practitioners. The result shows up in inventory: excessive stock you don’t need, and lost sales on the lines you care about.
Our packages were built to fix that, and they are grounded in peer-reviewed research, refined over a decade, and used by analysts worldwide. Here is what they implement:
Forecasts you can plan around. Every forecast translates directly into your decision, letting you set safety stock deliberately — hitting your target service level without drowning working capital in “just in case” inventory.
The hard cases, handled. Spare parts, components, and slow movers typically have intermittent demand that standard tools get badly wrong. Our methods were designed for precisely these profiles — the high-value, hard-to-predict items where better forecasting pays for itself fastest.
Demand explained, not just extrapolated. The models bring real drivers into forecasting — promotions, pricing, advertising spend, seasonality — so you understand why demand moves and can plan for what you are about to do, not only what already happened.
Glass-box, not black-box. Every model is transparent and interpretable. You can see the logic, explain the numbers to finance, and stand behind the plan.
Built to scale — and for short histories. Automated model selection picks the right approach for each SKU in split seconds, and works sensibly with the two or three years of data most businesses actually keep, where bigger models typically fail.
Proven, not promised. The methods are benchmarked against leading commercial and open-source tools on international forecasting competition data, and rest on a published methodology.
The toolkit
smooth (R & Python)
The core forecasting engine, built on the ADAM framework unifying exponential smoothing, ARIMA, and regression in one state-space model, with full probabilistic forecasts and dedicated handling of high frequency data and intermittent demand.
greybox (R & Python)
Model building and evaluation: bringing demand drivers into the model, selecting variables rigorously, and evaluating forecasts honestly.
legion (R)
Multivariate forecasting for when products, regions, or channels move together and shared dynamics carry information.
muse (C++)
Our research frontier: the Multiple Sources of Error state-space framework that extends where these models can go next.
All of it is open source, developed in the open at github.com/openforecast-org. For the technical documentation, worked examples, and function guides, see the smooth and greybox sections of this site.
The tools are the engine. The results come from applying them.
The packages are freely available, and that transparency is the point — you can inspect exactly what our methods do before you ever hire us. Our value is in the application: tuning the models to your data, streamlining them into your operational processes with the help of your ERP, correcting for stockout-censored sales, capturing your demand drivers, and translating the output into safety-stock and replenishment decisions your team can run every week.
If your forecasts drive real inventory decisions, we should talk.
Or read about our training first.