Can a cough reveal lung disease? Swaasa AI uses cough sound analysis to screen respiratory disorders

Artificial intelligence is steadily reshaping healthcare, and one of its most promising applications is emerging in respiratory disease screening. Researchers and healthcare startups are increasingly using AI systems to analyse cough sounds and breathing patterns to identify signs of lung disease at an early stage. Instead of relying entirely on expensive diagnostic equipment and hospital-based testing, some of these tools can work using the microphone on a smartphone, making respiratory screening faster and more accessible.

One such platform is Swaasa, an AI-powered respiratory screening tool developed by Hyderabad-based startup Salcit Technologies. The platform is designed to help detect respiratory conditions such as asthma, Chronic Obstructive Pulmonary Disease (COPD), bronchitis, pneumonia and tuberculosis by analysing a person’s cough sounds using artificial intelligence.

Traditionally, doctors use a test called spirometry to assess lung function. Spirometry measures how much air a person can inhale and exhale, and how quickly they can breathe out. While the test is considered the standard method for diagnosing many respiratory illnesses, it requires specialised equipment, trained technicians and proper clinical infrastructure. In many smaller towns, rural regions and primary healthcare centres, such facilities remain limited.


This is the gap that AI tools like Swaasa are attempting to address.

The system works in a relatively simple way. A patient coughs into a smartphone, tablet or laptop microphone, and the AI software analyses the sound recording. According to the company, the process can be completed within seconds and does not require any dedicated medical hardware.

The science behind the platform is based on the idea that respiratory diseases change the way air moves through the lungs and airways. These changes subtly affect the sound of a person’s cough. Certain conditions may create wheezing sounds, while others may alter the rhythm, sharpness or intensity of the cough. Many of these variations are difficult for the human ear to identify consistently, but AI systems can detect them by studying large amounts of audio data.

Swaasa uses machine learning, a form of AI that learns patterns from data, to compare a patient’s cough against thousands of cough recordings collected from people with different respiratory conditions. Over time, the system learns to recognise sound patterns linked to illnesses such as COPD and asthma.

The platform does not rely only on cough sounds. It also combines the audio analysis with other information such as symptoms, oxygen saturation levels and temperature readings to improve the accuracy of its assessment. Based on this analysis, the system provides an indication of whether the lungs appear normal or whether the patient may require further medical evaluation.

The platform is intended primarily as a screening and diagnostic support tool rather than a replacement for doctors or hospital testing. Its purpose is to help identify individuals who may be at risk and require additional clinical examination.

This approach could be particularly significant in countries like India, where respiratory diseases remain a major public health concern and access to specialised lung testing is uneven. In many healthcare settings, especially outside large urban hospitals, patients often face delays in diagnosis because pulmonary testing equipment is either unavailable or unaffordable.

AI-based cough analysis could help make early screening more widely accessible. Unlike conventional lung function tests, systems such as Swaasa do not require laboratory setups or highly trained operators. This creates the possibility of respiratory assessments being conducted in community health centres, small clinics and remote areas using only basic digital devices.

Healthcare professionals have also pointed out that respiratory illnesses are frequently underdiagnosed, especially at the primary care level. Symptoms such as persistent cough, wheezing or shortness of breath are sometimes mistaken for seasonal infections or allergies. Without proper screening tools, patients may not receive timely referrals for specialised care.

Swaasa aims to function as a first-level screening system that can assist healthcare workers in identifying potential respiratory problems earlier and more efficiently.

The platform has undergone multiple clinical validation studies involving thousands of participants. In a study presented at the European Respiratory Society Congress in 2024, researchers found that the system performed well in distinguishing healthy individuals from those with respiratory abnormalities.

Even so, the emergence of platforms like Swaasa demonstrates how artificial intelligence could expand access to respiratory healthcare. A process that once depended entirely on specialised hospital equipment may increasingly become available through simple digital tools capable of analysing something as ordinary as a cough.

(Disclaimer: The articles are not a substitute for professional medical advice, diagnosis, or treatment. Readers are advised to consult a qualified doctor, healthcare provider, or certified expert before making any changes to their diet, exercise routine, or lifestyle, especially if they have existing health conditions.)