Homomorphic Encryption
Compute directly on encrypted data — without ever decrypting it.
Homomorphic encryption (HE) is a class of cryptographic schemes that let you perform computations — addition, multiplication, and by extension arbitrary functions — directly on ciphertext, producing an encrypted result that decrypts to exactly what you’d get from running the same computation on the plaintext.
That means a server, cloud provider, or model host can process your data on your behalf without ever seeing it in the clear. Modern fully homomorphic encryption (FHE) schemes like CKKS and BFV have made this practical for targeted, high-value workloads, even though the computational overhead remains higher than plaintext processing — which is why PryvX pairs HE with other PETs (TEEs, MPC) rather than treating it as a one-size-fits-all default.
How It Works
Homomorphic Encryption in four steps
Encrypt
Data is encrypted client-side with a public key before it ever leaves the source.
Compute
The untrusted party runs the requested computation directly on the ciphertext, never seeing plaintext.
Decrypt
Only the data owner, holding the private key, decrypts the result.
Verify
The decrypted output matches exactly what plaintext computation would have produced.
Real-World Applications
Where HE shows up in production
Encrypted Credit Scoring
Banks score loan applicants on encrypted financial history without a third-party model provider ever seeing raw statements.
Confidential Health Analytics
Research institutions run population-level health queries on encrypted patient records.
Secure Cloud Outsourcing
Enterprises offload heavy computation to untrusted cloud infrastructure without exposing proprietary or regulated data.
Encrypted Search
Query an encrypted database — genomic, financial, or otherwise — and get matching results without revealing the query or the dataset.