Experimental free tool
Give Atlasly a UK location. It reads the site against official planning, heritage, ecology and ground datasets, works out what would genuinely move the project forward, then runs the next fortnight in front of you — checking every weather-dependent activity against a Google DeepMind WeatherNext 3 ensemble and adapting when a day will not work.
The run reads the same official datasets Atlas AI uses — Environment Agency flood zones, planning.data.gov.uk conservation areas and Article 4 directions, Historic England listings, Natural England designations, EA historic landfill and BGS ground screening. Where a dataset has no coverage, the run says so rather than reporting an all-clear.
Every field activity carries defensible limits — a UAV flight stands down above 9.5 m/s, a bat emergence survey needs 10 °C and light wind at dusk, a rig needs ground that three days of rain has not made impassable. WeatherNext 3 ensemble percentiles decide whether a day clears those limits, and how confidently.
When the next activity cannot happen, Atlasly says which limit stopped it, books it into the next window the ensemble supports, and spends the day on the work that can move. What you get at the end is a programme, with the blockers still named.