Privacy First

What happens to your lab file

When you upload a lab report — PDF or photo — it is read into server memory, processed by AI to extract marker names and values, then discarded. The original file is never written to disk, never stored in a database, and never sent to any third party in its original form. Only the structured data extracted from your report (marker names, values, reference ranges, flags, and collection date) is saved.

• Original file exists only in server memory during the ~15 seconds of processing
• No file storage bucket, no S3, no blob column in the database
• The AI model receives either extracted text or a rendered image of the page — not your original file
• After processing, the buffer is garbage-collected

What we store

We store only what is necessary to show you your results over time.

• Structured lab results: marker name, value, unit, reference range, flag
• Collection date and lab source name
• Your health profile: name, date of birth, sex, height, weight, blood group — used only to personalize analysis
• Health score and biological age calculations derived from your markers
• Account information: email address (hashed password via Supabase Auth)

What we do not store

• Your original lab report file
• Social Security numbers or government IDs
• Financial account numbers
• Any data not directly related to your health analysis

How your data is protected

• Encryption at rest: AES-256 on Amazon RDS — enabled at the storage layer and cannot be disabled
• Encryption in transit: TLS 1.2 or higher on all connections, including database connections using AWS RDS CA-signed certificates
• Row-level security: every database table has RLS policies enforced — your data is invisible to other users at the database layer
• Authentication: every API route verifies your Supabase JWT before any data is accessed or returned
• Ownership checks: every data read and delete operation verifies that the requesting user owns the record
• Audit log: all data access and deletion events are written to an audit log table

AI and your health data

Your lab marker values are processed by Claude (Anthropic) via AWS Bedrock to generate plain-English explanations, health summaries, and Doctor Prep briefs. This processing is covered by a Business Associate Agreement (BAA) with AWS, making it HIPAA-eligible.

• AI processing runs on AWS Bedrock — covered under our AWS BAA
• Marker values sent to the AI are structured data (name, value, unit) — not your original report
• AI responses are generated and returned — they are not stored by the AI provider
• We do not use your health data to train AI models
• Health Coach conversations are not stored server-side

Who can see your data

• Only you — your data is scoped to your authenticated account
• Family profiles you create — visible only to the account owner who created them
• Samvion Labs engineering staff — access is limited to aggregate, de-identified data for debugging. Individual health records are not accessed in normal operations
• No advertisers, insurers, employers, or data brokers ever receive your health data
• We do not sell your data. Ever.

Infrastructure

MyVitalSync runs on the following providers. Health data touches only providers covered by our AWS BAA.

• Amazon RDS (PostgreSQL) — primary database, HIPAA-eligible, covered under AWS BAA
• AWS Bedrock — AI processing, HIPAA-eligible, covered under AWS BAA
• Railway — web hosting, receives no health data in storage (stateless request routing only)
• Supabase — authentication only, no PHI stored
• Stripe — billing only, never receives health data
• Resend — transactional email, contains no PHI (no marker names or health data in any email)

Your rights

• Delete individual reports: from the Reports screen at any time
• Delete your account: from Settings — permanently removes all your data within 30 days
• Export your data: email privacy@myvitalsync.app
• Questions or concerns: privacy@myvitalsync.app