Legal · Test / validation period

Test-Period Disclaimer

OtoSensAI is in a regulatory test phase conducted under a CDSCO licence to test the software as a medical device. This notice explains the additional limits that apply during this period, on top of the standard Medical Disclaimer and the Terms that govern your participation.

OtoSensAI is being tested, not launched.

Voxmedai holds a licence to manufacture the software for the purpose of test under India's Medical Devices Rules, 2017. The tool has not received market approval as a diagnostic device.

This is not an early-access product, a beta, or a soft launch. It is a regulatory test phase with a defined scope, and every output produced during it is provisional.

1.What the test period is

The test period is the phase during which OtoSensAI is evaluated on real-world images, under a written protocol, to generate the validation evidence required for a future approval application.

Field	Detail
Licence	CDSCO Form MD-13, No. SW/MD/MD-13/2026/000004207
Type	Licence to manufacture for the purpose of clinical investigation, test, evaluation, examination, demonstration, or training
Issued	08 July 2026, by the Central Licensing Authority, New Delhi
Validity	Three years from the date of issue
Device class	Class B
Sites	C-3/387, Vishal Khand-3, Gomtinagar, Lucknow — 226010 Healthcity Vistaar Hospital, Sector-4, Opp Amar Shaheed Path, Vasant Khand Yojana, Gomtinagar Vistaar, Lucknow — 226010
Participants	Enrolled investigators and delegated staff named on the approved protocol
Cost	Free. There is no subscription and no billing during the test period.

The licence permits manufacture for test at the sites named above. It confers no right to sell, distribute, or supply the software, and the test period does not convert into a commercial launch. Any subscription arrangement arises only if and when market approval is granted, and is governed by the Terms & Conditions applicable at that time.

2.Results during the test period are investigational

Every output produced during this period is investigational data, generated to validate the model. That is the only purpose it may serve.

Outputs are not a diagnosis, not a clinical finding, and not authorised decision-support.
No output may be used to direct, delay, alter, or escalate the management of any patient.
No output may be communicated to a patient as a finding, an opinion, or reassurance.
No output may be recorded in a patient's medical record as a diagnostic result.

The tool must run parallel to care, never within it

Form your clinical impression from your own examination, before or independently of viewing any output.

Manage the patient exactly as you would if the tool did not exist.

An evaluation in which the investigational output influences care is not an evaluation. It is use of an unapproved medical device, and it also invalidates the data the test period exists to generate.

3.Accuracy figures are preliminary

Any accuracy figure discussed during this period reflects performance on training and validation data. Real-world accuracy may differ, and may differ substantially.

No accuracy figure may be relied upon as a guarantee.
No result may be treated as diagnostic.
Classification behaviour may change between model versions during validation — the same image may yield a different output at a later date.
Accuracy figures, interim findings, and images must not be used in any external communication, publication, or presentation without prior sponsor and regulatory review.

4.Your data helps validate the model

Collecting and analysing images is the purpose of this phase and the reason the licence exists.

Images are de-identified automatically at ingestion — metadata stripped, filenames discarded, and a pseudonymous identifier assigned — before storage.
Patient consent for AI-assisted analysis and for model development is obtained on the approved consent form, before capture, and may be withdrawn at any time.
The key linking study identifier to patient is held at site under the Principal Investigator's control, separately from the image store.
Access logs record site, investigator, device, and timestamp for every upload, and are retained as regulatory records under the Medical Devices Rules, 2017.
No personal or identifying information may be included in any image, filename, or note.

The limit of withdrawal

Withdrawal of consent removes an image from the dataset, stops all future use, and destroys the linking key. Once a de-identified image has been used to train a model, it cannot be extracted from that model. This is a technical characteristic of machine learning. Patients are told this at consent, before any image is taken.

5.The service will change, and may be interrupted

Features may be added, changed, or removed without notice as the model is refined.
The service may be temporarily unavailable during updates or maintenance.
Classification behaviour may change between model versions during validation.
The test period will be suspended immediately if CDSCO, the Ethics Committee, or our own safety review requires it.

6.Reporting issues

Because this is a validation phase, your reports matter. The obligation to detect, assess, and report device malfunctions and adverse events rests with us as manufacturer under the Medical Devices Rules, 2017; your part is to report promptly so that we can act.

Step	Timeline	Owner
Report any suspected misclassification with clinical consequence, malfunction, or adverse event	Same working day of becoming aware	Investigator
We acknowledge and open a vigilance record	Within 24 hours	Sponsor safety lead
Initial triage and severity assessment	Within 72 hours	Safety lead and Principal Investigator
Report to MvPI / CDSCO where reportable	Per Medical Devices Rules, 2017	Voxmedai
Root cause analysis and corrective action	Within 30 days of triage	Voxmedai
Suspension or field safety notice if warranted	Immediately on determination	Voxmedai, PI, and Ethics Committee

7.Reviewer escalation

Images and outputs are queued for review by ENT faculty affiliated with the evaluation, as a quality-assurance step. Review is a research activity: it does not create a doctor–patient relationship and does not constitute a formal medical opinion. Review is not real-time and may take several days.

If a Reviewer identifies an urgent finding

Where a Reviewer identifies a finding suggesting urgent or serious pathology, the Reviewer will notify the Principal Investigator at the originating site within one working day. The Principal Investigator is responsible for ensuring the patient is reviewed through the normal care pathway.

This is a patient-safety measure. It does not convert the review into a diagnosis, and it does not alter Section 2 — the investigational output still may not direct care.

8.Liability during the test period

The evaluation is conducted on a device that has not received market approval. The heightened, "as is" nature of an investigational product applies with full force during this period, and all limitations in the Medical Disclaimer and the applicable Terms continue to apply.

We do not exclude or limit our liability for death or personal injury caused by our negligence, for fraud, or for any other liability that cannot lawfully be excluded.
Subject to the above, and to the maximum extent permitted by law, we accept no liability for any clinical decision made in reliance on an investigational output.
Clinical trial and product liability insurance is maintained for the evaluation. Details are available on request.
Compensation for study-related injury is governed by the approved protocol and applicable rules.

9.Everything else still applies

This notice adds to the Medical Disclaimer, the Medical Image Guidelines, and the Terms governing your participation. It does not replace any of them.