What the application actually does

AI Health Chatbot

A conversation where you can ask anything about a symptom, a medical term or a general health topic.

What it does

• Answers in short headed sections: what this may mean, general care, when to see a doctor
• Remembers the last few messages so follow-up questions make sense
• Keeps each conversation in a list you can return to or delete
• Replies in English, Hindi or Marathi

Where it stops

• It will not name the condition you have
• It will not recommend a medicine or a dose
• There is a per-user limit on requests each hour and each day

Medical Report Summaries

Upload a report you did not understand and get it explained in plain language.

What it does

• Reads text-based PDF files and plain text files
• Explains test names, values and units using only what is printed on your document
• Shows you the exact text it read, so you can check nothing was invented
• Lets you ask follow-up questions about that specific report

Where it stops

• Photographs and scanned images are stored but cannot be read — there is no OCR
• It never invents a reference range that your document does not contain
• It does not diagnose from your results or suggest treatment

Symptom Guidance

A structured form that turns "I feel unwell" into something specific enough to be useful.

What it does

• Pick from fourteen common symptoms or describe your own
• Add duration, severity, age and any extra detail
• Returns general information, self-care measures, what to monitor and when to see a doctor

Where it stops

• The result is labelled general guidance, never a diagnosis
• It does not rank possible diseases or give probabilities

Emergency Detection

Plain program logic that recognises red flags before the AI is involved at all.

What it does

• Eight rules covering chest pain, breathing difficulty, loss of consciousness, severe bleeding, seizures, severe allergic reactions, sudden neurological symptoms and thoughts of self-harm
• Checks both the symptoms you selected and anything you typed
• Shows emergency guidance with local numbers instead of an AI answer
• Rejects any AI response that would downplay a detected red flag

Where it stops

• It matches phrases, so unusual wording can be missed. It is a safety net, not triage
• It never decides that something is safe — only that something looks urgent

Medicine Information

A curated reference of common medicines, with the database as the source of truth.

What it does

• Search by brand name, generic name or what it is used for
• Uses, side effects, precautions and warnings for each medicine
• Marks clearly whether a prescription is required
• Names the kind of source the entry should be checked against
• Optionally rewrites the same stored information in simpler words

Where it stops

• No dosage instructions of any kind
• No drug interaction checking
• The AI may only rephrase the stored entry, never add to it

Medicine Reminders

A simple schedule of what you take and when.

What it does

• Pick a medicine from the catalogue or type any name
• Set a time, a frequency, a start date and an optional end date
• See what is due today as soon as you open the page

Where it stops

• There are no notifications, emails or text messages — reminders appear in the app only
• This feature does not use the AI at all

Health History

A small personal record: notes, conditions and measurements over time.

What it does

• Record weight, blood pressure, heart rate, blood glucose or temperature
• See any one metric as a trend chart, with the same figures in a table
• Keep notes and mark conditions active or resolved

Where it stops

• It is a personal log, not a medical record system
• Values are not interpreted for you

Three Languages

English, Hindi and Marathi across the entire interface.

What it does

• Every label, button, warning and error message is translated
• Your language is stored on your account and remembered
• The assistant answers in the language you chose

Where it stops

• Seeded medicine and emergency text is currently stored in English only

Privacy and Security

Health data is treated as sensitive, and the boundaries are enforced in code.

What it does

• Reports are stored outside anything the web server will serve
• Every record is checked for ownership before it is shown
• Deleting a report deletes the stored file as well
• Administrators see row counts, never your conversations, reports or records
• The AI key stays on the server and is never sent to your browser

Where it stops

• Your question and the text of a report you upload are sent to the AI provider to be explained