The patent (granted in US, IN, IL; imminent grant in EP for 2026; pending in CN) and a set of supporting R&D-stage assets: a high-performance Rust implementation of the prediction and training model, several color science methods and optimizer, an internal research dashboard for experiment tracking and prediction analysis, supporting data-preparation tooling, the public presentation site at colorfit.ai, an interactive demo environment for walking through job selection and parameter adjustment on the Lab curve. All software is production-quality R&D — the acquirer takes it to production in their own stack. Documentation covers the algorithm, its mathematical and physical foundation, and integration paths into press, QC, or color-management software.

ColorFit takes job-specific inputs — substrate, ink, and press conditions — and predicts the print settings that hit the ΔE target before makeready begins. The press becomes a predictable control point, not a correction step.

For converters, the impact is measurable: less rework, faster makeready, more consistent print quality, and reduced material waste and downtime — EUR 200,000 in operational savings per press per year (see the Savings section). The result is first-time-right color matching on press.

Traditional color management stops at profiling and proofing, and inspection systems only detect drift after it happens. The ColorFit Method predicts the correct print parameters before makeready and adjusts them during the run, so target color is achieved on the first pull rather than chased by trial and error.

No. The inference runtime is lightweight and the algorithm physically consistent — rooted in proven color science rather than large-scale machine learning, and small enough to run on press-side hardware. The patent claims describe the prediction approach, not a deployment architecture, so the acquirer is free to host inference and training wherever fits their platform.

A typical recommendation is to run training in the cloud, where measurement data from several presses can be aggregated and training jobs orchestrated across them. Inference then sits on-press, on-prem, or wherever latency and integration demand.

The patent claims cover the printing method itself — the scope is not limited to flexography or gravure. Validation to date has come from internal and external studies in flexo and gravure.

Job target data (target Lab or spectral reflectance curves), the allowable range of dynamic press variables, and inline spectral measurements. The model also incorporates ink recipe data, substrate properties, and pressroom conditions as they become available.

The ColorFit Method is co-developed by ColorFit GbR and Siegwerk Druckfarben AG & Co. KGaA. Both parties are joint owners of the patent, in equal shares.

Press manufacturers, color-management vendors, and inspection or quality-control providers who want to extend their platforms from observation into closed-loop control. The acquisition fits parties that can integrate the method into their existing product line.

The inventors and representatives are listed in the Contact section. Frank Dauth (Siegwerk) and Markus Unkel (ColorFit) handle inquiries directly.