It's not just Molekule, consumer air quality products in general are full of dubious junk. This study [0] fairly comprehensively compared the Foobot air qualit

It's not just Molekule, consumer air quality products in general are full of dubious junk. This study [0] fairly comprehensively compared the Foobot air quality monitor to a research-grade sensor setup and found that it severely underestimated the temperature (one of the easiest things to get right in my experience), underestimated tVOC, had low correlation on its CO2 output (i.e. the CO2 output was pure garbage) and overestimated PM2.5. I wish this was unique to Foobot. The problem seems to be that these companies take a bunch of cheap, low-quality sensors (e.g. using inferred CO2 rather than an NDIR sensor or saving cost by not properly calibrating the sensors), put them in a premium-looking (and priced) package and market them as if they were premium products, because consumers won't know any better. The worst part is that high-quality sensors really aren't that expensive. A bundle of Sensirion tVOC, CO2, RH/T and particulate sensors can be had for under $200 and with proper calibration they work great. Hook them to an ESP32 or something and put them in a case and you've basically got a Foobot. [0]: https://www.j-sens-sens-syst.net/7/373/2018/ https://www.j-sens-sens-syst.net/7/373/2018/