The whole dossier for one land parcel in a single call: the parcel core (location, area, land use, plan designation), all thirteen enrichment layers (flood risk, heritage listing, landslide risk, subsurface: mining terrains and major groundwater reservoirs, nuisance surroundings, public-transport access, general-plan zoning, buildings on the parcel, recent building activity, agricultural-land eligibility, official land-use & soil-quality classification with its re-designation consequences where the county publishes it, nature: nearby forest and protected areas, roads: geometric road-access evidence measured from carriageway centrelines, which is not a determination of legal access), the parcel's transaction history (newest first, up to 20), a local price context (median zł/m² per property type for the county and the gmina over the last 12 months) and a municipal context (a headline demographic/economic subset plus upcoming-infrastructure signals for the gmina). Address it by a full cadastral id in the natural '/' form ('142907_2.0014.342/5'), the URL-safe '-' form, or the internal UUID from a search or resolve result. Each section carries its own state, shown explicitly: covered = a definitive result; covered_no_data = the parcel was checked and nothing was found (still billed); not_covered = outside our data (refunded); not_computed = a live computation could not finish in time (refunded — the rest of the report still returns, so a report can be partial). The context sections instead use full / low_sample / suppressed / no_data. Covered sections carry their details under the summary line: individual heritage entries (name, category, and whether the parcel lies within a protected area or a listed object stands on or at it), flood scenarios, building-activity cases, the distance to every nuisance category including high- and extra-high-voltage overhead power lines (or none within the search radius), and the names of protected areas, groundwater reservoirs and mining terrains. Long lists are capped with a count of the rest. The report also carries a nearby designation signal: the plan designations recorded in RCN deed data for neighbouring land transactions (with their count, radius and period). It describes the neighbourhood, not this parcel, and is not taken from the planning register; confirm the parcel's status with an extract from the local plan (MPZP) or a zoning decision. It is never a statement that a parcel can or cannot be built on. A parcel shape section describes the geometry of the cadastral outline only: area and perimeter, the minimum bounding rectangle (width and length), the widest circle that fits inside, compactness, a shape class, the side lengths of a simple outline, and the area that remains at least 4 m and 3 m from every boundary. It is a geometric illustration, not a building-permit determination, and says nothing on whether the parcel may be built on or divided; the local plan or a zoning decision, road access and setbacks decide. Its states are full / no_data, like the other context sections. The buildings section additionally reports how many of the buildings could be given a construction-age estimate from building-permit records. That is an ESTIMATE with an interval, never a registry construction date, and the records only start in 2016, so for most buildings the answer is that the year could not be established — which is stated rather than omitted. Prefer this over calling the per-layer parcel tools one by one — it is one call at a flat price and never costs more than the sum of its parts. Use resolve_parcel first when you only have an address, a coordinate, or a 'locality + number'. Costs 45 API tokens. Billing is by outcome (see the billing line on the response): a parcel that cannot be resolved is fully refunded; a resolved parcel where no layer had data is billed only the core floor (1 token) with the rest refunded; a resolved parcel with at least one covered layer is billed in full. Requests are rate-limited per API key per minute, so parallel agents sharing one key share the same limit. On a "Too many requests" response, wait the number of seconds it gives and retry sequentially.