This work describes the development of a revolutionary new class of universal perception-based sensor capable of diagnosing a wide range of diseases, and compatible with low-cost mass production. This sensor takes the form of an array of fluorescent, environmentally sensitive single walled carbon nanotubes (SWCNTs), resembling a conventional two dimensional QR code. Each group of pixels of the QR code is formed from carbon nanotubes with different surface functionalisation and different environmental sensitivity. Addition of an analyte such as serum causes each pixel to change its fluorescence, generating a unique spectral ‘fingerprint’ characteristic for each disease. A machine-learning approach is then trained to identify the disease based on their corresponding fingerprint.The basic principle of this technology has been previously described by the Heller Lab for the detection of ovarian cancer, but exclusively in a microplate-based format which requires extensive liquid handling. Translating this technology into a microarray format offers a number of significant advantages, including easier handling, lower production cost, less training required for operation, simpler read-out system that can be based on mobile phone devices and the ability to simultaneously screen for multiple diseases. In this way, this technology advances beyond the current state of the art of SWCNT-based sensors. In a microarray format, hundreds of nanotube assemblies can be analysed simultaneously.By creating an array of sufficient complexity, we hope to remove the need to tailor an array for a specific disease that act as a universal platform for diagnosis of a range of different diseases. This format also significantly lowers the number of necessary sample handling steps, reducing the potential for user error and the amount of training necessary to employ the sensor. Figure 1