Embodied carbon in building is the carbon dioxide emissions associated with materials and construction processes throughout the whole lifecycle of a building. The carbon footprint of existing building can be used as an imperative indicator or reference for future carbon taxes and house building to meet climate goals. This paper presents a concept for quickly deriving potential material inventories from the building stock and estimating the CO2 footprint over the building's entire life cycle. The web application will grab the geometry information from CityGML model, and machine learning model is deployed to extract the cross-sectional area of walls from floor plan. As long as the obtained substance and material inventories are combined with the ÖKOBAUDAT material database, the embodied carbon emissions of the building can be determined.