h33 0.1.0

H33 Post-Quantum Encryption SDK — Kyber + AES-256-GCM hybrid encryption, FHE biometrics, and more

h33 — Post-Quantum Encryption SDK for Python

The official Python SDK for H33, the post-quantum encryption platform delivering 1.6M authentications/sec with lattice-based cryptography.

Install

pip install h33

Quick Start

from h33 import H33Client

client = H33Client(api_key="h33_live_...", base_url="https://api.h33.ai")

# Post-quantum storage encryption (Kyber + AES-256-GCM)
result = client.encrypt_blob(b"sensitive data")
ciphertext = result["ciphertext_b64"]
key_id = result["key_id"]

# Decrypt
plaintext = client.decrypt_blob(ciphertext, key_id)
assert plaintext == b"sensitive data"

Storage Encryption

Encrypt data at rest with hybrid post-quantum encryption (ML-KEM/Kyber + AES-256-GCM):

from h33 import StorageClient

storage = StorageClient(api_key="h33_live_...")

# Encrypt with optional authenticated data
result = storage.encrypt(b"patient record", aad=b"user:12345")
print(result["key_id"])  # Kyber key ID for decryption

# Decrypt
plaintext = storage.decrypt(result["ciphertext_b64"], result["key_id"], aad=b"user:12345")

# Field-level encryption with automatic sensitivity classification
fields = storage.encrypt_fields([
    {"name": "ssn", "value": "123-45-6789"},
    {"name": "name", "value": "Jane Doe"},
    {"name": "email", "value": "[e-mail]"},
])
# Each field gets a sensitivity level (Confidential, Internal, etc.)

# Key rotation — re-encrypt under the current active key
rotated = storage.rotate(result["ciphertext_b64"], result["key_id"])

Client-Side FHE (Homomorphic Encryption)

Encrypt biometric embeddings on the client so the server never sees plaintext:

from h33 import FheEncryptor

# Load public key from H33 API
enc = FheEncryptor.from_public_key("public_key.json", embedding_dim=128)

# Encrypt a 128-dimensional face embedding
embedding = [0.1, 0.2, 0.03, ...]  # from your ML model
ciphertext_json = enc.encrypt_embedding(embedding)

# Send ciphertext to H33 for matching — server operates on encrypted data

Requires the h33-fhe-client native library. Build from source:

cd h33-fhe-client && cargo build --release

API Reference

H33Client

Method	Description
health()	Check API status
send_code(phone)	Send SMS OTP for authentication
verify_code(phone, code)	Verify OTP and get session
create_key(label)	Create a new API key
list_keys()	List all API keys
revoke_key(key_id)	Revoke an API key
encrypt_blob(plaintext, aad)	PQ-encrypt arbitrary data
decrypt_blob(ciphertext_b64, key_id, aad)	Decrypt PQ-encrypted data
encrypt_fields(fields)	Field-level PQ encryption
rotate_encryption(ciphertext_b64, old_key_id, aad)	Re-encrypt under current key
fhe_schemes()	List available FHE schemes
noise_estimate(n, t, q_bits, operations)	Estimate FHE noise budget

StorageClient

Convenience wrapper for encryption-only workflows:

Method	Description
encrypt(plaintext, aad)	Encrypt bytes
decrypt(ciphertext_b64, key_id, aad)	Decrypt bytes
encrypt_fields(fields)	Field-level encryption
rotate(ciphertext_b64, old_key_id, aad)	Key rotation

Platform

H33 is the world's fastest post-quantum authentication platform:

• 1.6M auth/sec on AWS Graviton4 (96 vCPUs)
• ~42 microseconds per authentication
• Kyber + AES-256-GCM hybrid encryption for data at rest
• BFV Fully Homomorphic Encryption for biometric matching on encrypted data
• Dilithium digital signatures for attestation
• Zero-knowledge proofs for privacy-preserving verification

Learn more at h33.ai

Documentation

• H33 Platform
• API Documentation
• FHE Overview
• Post-Quantum Cryptography
• White Paper
• Benchmarks

License

MIT — see LICENSE