The thing almost everyone already does
You get a PDF from a lab portal. The ranges are tiny. The abbreviations are hostile. You want a plain-language read, so you paste the panel into ChatGPT, Claude, or Gemini.
That is a reasonable impulse. It is also a quiet decision to hand a copy of your medical history to a consumer chatbot — often with an account, a retention policy you have not read, and a training default you did not set.
DoctorVault exists for the people who will not keep doing that, and who still want a specialist-style second look at their file.
Convenience is not custody
A general chatbot is optimized to answer the next prompt. It is not a vault. Pasted labs sit in that product's logs, project files, or conversation history under their terms.
DoctorVault is built the other way around. Records are encrypted in your browser and synced to storage you own — Google Drive or an S3-compatible bucket. Our servers ship the app shell. They do not receive plaintext health records, chats, files, passphrases, or API keys. There is no endpoint that accepts them. You can verify that yourself on the security page.
Context beats a one-off paste
Pasting one CBC into a chatbot answers one PDF. It does not remember last year's thyroid labs, the lisinopril you already take, or the imaging report you uploaded in March.
In DoctorVault, structured records (labs, medications, conditions, files) live in the vault. When you ask the panel a question, specialists retrieve relevant history and a GP synthesizes one answer. That is the product: persistent context plus a debate you can read — not a disposable chat window.
What we will not pretend
Three things people get wrong, including us if we get sloppy in marketing:
- This is not a doctor. DoctorVault is a personal health information tool. It does not diagnose, prescribe, or replace a licensed clinician.
- This is not HIPAA-certified. Architecture reduces what we can see. It does not make your LLM vendor a covered entity, and we do not sign BAAs.
- Chat is not fully private. When you consult the panel, relevant excerpts go from your browser to the AI provider you configured, under your key. We are not in that path — they are. Read what still leaves your device.
Start with the demo, not a bucket
You do not need S3 on day one. Open the demo vault, add a couple of sample records (not real labs if you are just kicking the tires), ask one question, and watch the panel work. The demo lives in this tab; close it and it is gone.
When you want a real home for the file, connect Google Drive or S3 and keep the same passphrase model. Recovery is on you — we cannot reset a passphrase we never hold.
Try it in the demo vault
No account. Nothing is saved when you close the tab. Add a couple of sample records, ask the panel one question, then read how to verify that our servers never see the contents.