New unique technique for diagnosing diseases of the respiratory system will speed up and simplify the process diagnoses in medical facilities! The project aims to improve the quality of human respiratory system diagnosis by developing an expert pulmonology diagnostics system based on modern methods of analysis of macromolecular structure of moisture in exhaled air.Our proposed approach is innovative, there are only indirect competitors, whose methods are more time consuming and costly.Target users of system are doctors and specialized diagnosis centers operators, who have equipment for receiving and processing samples of condensate moisture in exhaled air.We offer physicians familiar with the system and rid yourself of the routine - to put the preliminary diagnosis. Just ask your patient to exhale in a special container and run the system.The main purpose of “Spectrex” is to determine the state of human respiratory system using computer models of the diagnostic features of respiratory system. This allows you to assess the condition of the respiratory system of the patient as normal or reveal its changes is an indicator of a particular disease. Prototype Spectrex has already been able to identify bronchitis and pneumonia with an accuracy above 75%, the system is constantly updated with new diagnostic data. Using machine learning algorithms (neural networks, non-linear classifiers) allows for flexibility of the constantly increasing number of recognized diseases and their recognition accuracy. Necessary is no longer at the first suspicion of the disease to spend tedious, costly and often harmful survey. Compact diagnostic equipment will carry out surveys in the therapeutic and prophylactic purposes in kindergartens, schools, maternity homes and nursing homes, as well as in the field of sports medicine.We consider possibility of creating the mobile application with mobile spectrometer.To build a system were used software packages Eclipse, Netbeans (Java and C), MySQL.