We present MINT (Multi-participant Interactive Trading), a platform for experimental economics research designed to run controlled multi-participant trading sessions with flexible role assignments. The system features an event-driven architecture, session pool management, and seamless integration of human participants with algorithmic traders. It implements lazy market creation to reduce resource consumption, real-time WebSocket communication, and comprehensive data collection across more than 40 parameters. Authentication options include integration with online recruitment tools such as Prolific, and the software is released as open-source to foster adoption within the research community.

Using MINT, we study strategic behavior and information transmission in financial markets through a controlled trading experiment. In each market, algorithmic noise traders provide liquidity while an algorithmic informed trader must execute a target order. A human participant trades in the same market with the objective of making profits. Experimental treatments vary along key dimensions, including the size of the order to be executed, whether the order is a buy or sell, and the set of trading actions available to the informed trader (restricted to aggressive orders or allowed to use both aggressive and passive strategies). This framework allows us to investigate how market participants respond to execution pressure, how information is revealed through order flow, and how market design affects trading outcomes.