Practical TFHE for SaaS

Run TFHE workloads in production without exposing data-in-use.

Lattica enables TFHE-powered applications that involve complex decision logic in partnership with Sunscreen. Anything from genetic imputation to smart contracts.

image-1

Two Paths for Encrypted Computation
CKKS vs. TFHE

CKKS

  • Ideal for AI Workload.
  • Excels for data-heavy tasks.
  • Parallelized operations on multiple values.
  • GPU‑accelerated.

FHE Applications

Encrypted DB
AI Inference
Smart Contracts

FHE Software Stack

Algorithms

CKKS
TFHE

Core Building Blocks

Polynomial arithmetic
KeySwitch
Bootstrap

TFHE

  • Ideal for secure voting or blockchain logic.
  • Excels at exact computation.
  • Bootstrapping by design.
  • CPU-based.

Under the Hood: Sunscreen and Lattica

Lattica

Deploys and Operates

  • Orchestration and workflow automation
  • Query Clients
  • Computer resource allocation
  • Management and monitoring

Sunscreen Compiler

Compile your program

arrow
arrow
arrow

Lattica
Management Client

  • Create a key set

  • Encrypt

  • Decrypt

arrow

Deploy Compiled Program

arrow

Deploy Evaluation Key

arrow

Encrypted Query

arrow

Encrypted Response

TFHE Engine
(Cloud)

Sunscreen

The TFHE Engine

  • Encrypted computation at the language and compiler level: TFHE Compiler (Parasol)
  • TFHE runtime for encrypted programs
  • See Sunscreen demos...

TFHE and CKKS are commonly used names for a family of FHE algorithms