Trial matching, without exposing your medical data
Active Trials
Registered clinical trials
Protected Patients
Privacy-preserving registrations
Data Privacy
100%
End-to-end encrypted
One platform, two layers of privacy
GenLayer AI Advisor
An AegisCareAdvisor Intelligent Contract explains eligibility, recommends trials, and validates registrations against live ICD-10 data — all through LLM-backed consensus.
Optimistic Democracy
Every advisor result is proposed by a leader and verified by independent validators with equivalence rules, so non-deterministic LLM output settles on-chain trustlessly.
End-to-End Encryption
Medical data is encrypted before leaving your browser and stays encrypted throughout the entire matching process.
FHE-Powered Matching
Eligibility is computed on encrypted data using Zama FHEVM — no plaintext exposure at any point.
Strict Privacy Boundary
The advisor only ever sees anonymized inputs — age buckets, condition categories, and PII-screened summaries. Raw or encrypted PHI never leaves the FHE layer.
Private Results
Only you can decrypt your eligibility results, using an EIP-712 signature from your own private key.
How it works
For Patients
- 1Register with encrypted medical data (age, gender, BMI, conditions)
- 2Browse available clinical trials
- 3Check eligibility — computed on encrypted data on-chain
- 4Decrypt your result with an EIP-712 signature (only you see it)
For Trial Sponsors
- 1Create a trial with encrypted eligibility criteria
- 2Set age range, gender requirements, BMI limits, condition codes
- 3Patients check eligibility without revealing their data
- 4Privacy guaranteed — you never see patient medical data
A dual-chain architecture
FHE handles the confidential math; GenLayer handles the judgment. Each layer does the one thing the other fundamentally cannot.
FHE Matching Layer
Zama fhEVM · Solidity · Sepolia
Computes eligibility entirely on encrypted data. Medical values and trial criteria never appear in plaintext on-chain.
- ▸ Encrypted patient registration & trial criteria
- ▸ FHE comparisons on encrypted euint values
- ▸ EIP-712 private decryption by the patient only
AI Advisor Layer
GenLayer · Python · StudioNet
Runs LLM-backed logic with leader/validator consensus on anonymized inputs only. Explains, recommends, validates, and checks eligibility.
- ▸ generate_explanation — plain-language result
- ▸ recommend_trials — best 1–3 matches
- ▸ validate_trial — checks live ICD-10 reference
- ▸ check_eligibility — PII-screened summary
See AegisCare in action
Master AegisCare
Everything you need to know about privacy-preserving clinical trial matching — 10 sections, 50+ FAQ answers, step-by-step tutorials.