Habitat connectivity — how well a fragmented landscape lets species move between patches of suitable habitat — is a central input to conservation planning and environmental impact assessment. The established tool in the field, Conefor, computes the Probability of Connectivity (PC) index and its per-patch decomposition (dPC), but its core algorithm becomes intractable above about 1 000 habitat patches: a single such analysis can take a week of computing time, and larger landscapes cannot be analysed at all. ekokrati.graph computes the same connectivity metrics with modern sparse-graph algorithms whose results are mathematically exact — validated against Conefor — and which extend exact analysis to around 10 000 patches. Beyond that, a hierarchical spatial-decomposition method targets national scale (up to ~250 000 patches) with quantified, ecologically validated bounds on the approximation error — a capability no existing tool provides. This allocation funds a first phase on Dardel's CPU nodes, where the software runs unmodified: validating the method at scale, producing the first measured compute-cost figures, and running one representative Swedish landscape through the full pipeline. The result is open scientific software and a peer-reviewable methodology that lets ecologists, agencies, and EIA practitioners analyse connectivity at the scale at which conservation decisions are actually made. The measured results from this phase will justify a later, larger allocation for the full national programme.