Carbon dioxide enhanced oil recovery (CO2-EOR) has been widely used to improve production from mature oil fields around the world. To be effective, the injected gas and reservoir oil must develop miscibility, which generally requires prolonged contact between the two phases while in relative motion. Thus, identifying whether miscibility is possible is crucial for determining the feasibility of such EOR projects. The current industry-standard method of characterization, the slim-tube, requires weeks of analysis, while alternative methods are unable to infer all routes to miscibility, producing significant overestimates in required pressures. Microfluidic devices have the potential to simplify and speed up the analysis by offering high levels of fluid control and excellent visualization. Recently, high-pressure microfluidic devices etched into glass and exploiting crude oil's natural fluorescence have been successfully demonstrated. Here we focus on designing a microfluidic channel for identifying the development of miscibility. We prove its accuracy for a known ternary fluid system that mimics the true oil-gas system and can be manipulated at room temperature and pressure. Our chip consists of a single channel with several inline pocket structures. The chip is initially flooded with one phase before a second phase is injected via a flow-rate-controlled pump. The first phase is then rapidly displaced in the primary channel, but small samples are retained within the pockets. Over time, these trapped droplets can be observed as they interact with the continuously flowing second phase. When the fluid concentrations meet the conditions for development of miscibility, a dramatic and visually observable change in behavior occurs, allowing for characterization within 2 h.