Cough Into A Smartphone: How An AIIMS-Validated AI Software Is Screening For COPD In Under 10 Minutes

A 10-second cough recorded on a smartphone may soon help doctors screen for chronic respiratory diseases in settings where advanced diagnostic machines are unavailable.

Developed by Bengaluru-based Salcit Technologies, the software, Swaasa has been validated in multiple hospital settings, including by researchers at the All India Institute of Medical Sciences (AIIMS). Classified as a “software as a medical device”, it requires no additional hardware beyond a standard smartphone.

The premise is simple but scientifically layered: cough sounds carry disease-specific acoustic signatures. By analysing these signatures using an AI-based algorithm, the software attempts to detect underlying respiratory abnormalities.

It features among the 10 high-impact AI tools developed in India, as listed in the official compendium of the ongoing AI Impact Summit 2026.

Narayana Rao Sripada, founder of Salcit Technologies, said the person is asked to take a breath and cough into the mobile phone microphone for up to 10 seconds. “In that time, a person can take three to four breaths and cough. The microphone captures the signal and the algorithm decodes it,” he said.

Within roughly eight minutes, the software generates a structured output. It first categorises the recording as normal or abnormal. If abnormal, it further classifies the pattern as obstructive, indicating airway narrowing as seen in asthma or chronic obstructive pulmonary disease (COPD), or restrictive, where the lungs are not expanding adequately. It also flags whether asthma, COPD or a pulmonary process is “likely”, making clear that the output is indicative, not confirmatory.

Sripada mentioned that the software is not a replacement for spirometry — the gold standard lung function test — but a screening and decision-support tool, particularly valuable in primary and secondary health facilities where spirometers are often unavailable.

A Pulmonary Function Test (PFT) or spirometry test in India generally costs between ₹600 and ₹1,500, with an average, affordable price often around ₹750 to ₹1,000. Costs vary by city, diagnostic center, and whether it is a basic test or includes advanced measurements like DLCO, which can increase the price. 

That was precisely the context in which AIIMS researchers evaluated it.

Dr Harshal Ramesh Salve, Additional Professor at AIIMS’ Centre for Community Medicine, led a validation and feasibility study in a secondary-care setting at Ballabhgarh. The team recruited 460 COPD patients and compared the cough-based results against spirometry and chest X-ray findings.

“We found moderate correlation overall, and in severe disease, the correlation was excellent,” Dr Salve told Outlook. He added that the software can be useful in public health facilities where spirometry is not routinely available. “Around 20–25 per cent of public health facilities do not have spirometry. In such settings, this can aid doctors in clinical decision-making.”

Beyond the AIIMS study, Salcit Technologies says it has conducted multiple validation exercises across India, including at Christian Medical College Vellore and Apollo Hospitals, evaluating close to 12,000 subjects. 

According to Sriapada, the software is about 90 per cent accurate in distinguishing normal from abnormal cases, 85–87 per cent accurate in detecting asthma and COPD, and around 79 per cent accurate in identifying pulmonary processes.

Since deployment, more than 4,50,000 assessments have been conducted on the platform,Sripada said. It holds a Class B medical device (low-to-moderate risk) licence from Indian regulatory authorities. 

The software is currently intended for use by doctors and trained clinical personnel, not for patient self-diagnosis. Clinicians enter basic details such as symptoms, blood pressure and medical history before recording the cough sample. Unlike some community-level AI tools that depend on extensive questionnaires, this system relies primarily on acoustic analysis, which developers say reduces language barriers.

Deployment is underway through partnerships with NGOs, private health providers and pilot projects with state governments. The software has been used in parts of Karnataka, Maharashtra, Uttar Pradesh, Bihar, Telangana, Andhra Pradesh and Odisha, either through non-profit organisations or health networks. In Andhra Pradesh, it is part of a state-led innovation pilot across selected districts.

The technology has been in development since 2017. Rao described the journey as one involving years of proof-of-concept work, algorithm refinement, clinical validation and regulatory approval. “Initially, it was difficult to convince people that cough can be used as a marker,” he said. “But with evidence generated through multiple validations, acceptance is increasing.”

AIIMS researchers echo that the technology should be seen as an adjunct rather than a replacement. “We are not replacing spirometry,” Dr Salve said. “But where spirometry is unavailable, this can help doctors decide who needs further evaluation.”

With respiratory diseases often underdiagnosed and detected late, especially outside tertiary hospitals, both the developers and public health researchers argue that scalable screening tools could help bridge a long-standing diagnostic gap,  starting with something as ordinary as a cough.