The rise of order-driven markets in recent years created considerable challenges for researchers, who have to cope with extremely large amounts of data produced daily by the markets. It is our goal to spare academic researchers the tedious task of technical pre-processing of the data and thus enable them to focus on economic research. Our order book reconstruction system LOBSTER is based on the generalized order-processing algorithm common for majority of order-driven markets and so it is easily adjustable to process data produced in any order-driven market. Efficient data structures result in very convincing performance, with large datasets produced on the fly. Currently LOBSTER uses ITCH data from NASDAQ to accurately replicate the limit order book for any NASDAQ-traded stock to any desired level. Data from more venues will be added in the future. The system is accessible via the Internet, which makes it very convenient for researchers from around the world.