Skip to content
Theos Quantum Θ mark

The AutoPQC platform

Automated discovery and migration of post-quantum cryptography.

Four modules take an enterprise estate from unknown cryptographic exposure to a benchmarked, standards-anchored migration program.

Module 01

Discovery Engine

You cannot migrate what you cannot see. The Discovery Engine scans an organization's estate without agents on the critical path: AST-based static analysis reads Python codebases at the syntax-tree level, JCA signature-pattern analysis covers Java, and X.509 certificate and TLS configuration parsing maps cryptography in transit. The output is a complete Cryptographic Bill of Materials (CBOM) — a living inventory of every algorithm, key, certificate, and protocol in use, with location and role.

  • AST-based static analysis for Python codebases
  • JCA signature-pattern analysis for Java
  • X.509 certificate & TLS configuration parsing
  • Complete Cryptographic Bill of Materials (CBOM) output
AST analysisJCA patternsX.509 / TLSCBOM
Python ASTJava JCAX.509 / TLSCBOM

Module 02

Migration Mapping Engine

Post-quantum migration is not one-size-fits-all. An RSA key used for key establishment needs a different successor than an ECDSA key used for signatures. The Mapping Engine classifies each asset by role — key establishment, digital signature, data-at-rest protection — and maps it to the correct NIST standard: ML-KEM (FIPS 203) for key establishment, ML-DSA (FIPS 204) and FALCON for digital signatures, and SLH-DSA (FIPS 205) for hash-based signatures. Every mapping is standards-traceable.

  • Role-aware classification (key establishment, signatures, data-at-rest)
  • ML-KEM (FIPS 203) mapping for key establishment
  • ML-DSA (FIPS 204), FALCON mapping for digital signatures
  • SLH-DSA (FIPS 205) for hash-based signature use cases
ML-KEM · FIPS 203ML-DSA · FIPS 204SLH-DSA · FIPS 205FALCON
CBOM assetRole classifierNIST catalogPQC target

Module 03

AI Recommendation & Prioritization

A multi-year migration lives or dies by sequencing. Our Random Forest model ranks every discovered asset by risk and urgency, so teams migrate what matters first. Each recommendation ships with explainable per-asset reasoning — the factors that drove the ranking are visible, auditable, and defensible in front of a risk committee. The model currently achieves 95.2% classification accuracy with a 0.94 macro-F1 score.

  • Random Forest risk & urgency ranking
  • 95.2% classification accuracy, 0.94 macro-F1
  • Explainable per-asset reasoning for every recommendation
  • Estate-level risk scoring
Random Forest95.2% accuracy0.94 macro-F1Explainable AI
Asset featuresRandom ForestRisk rankingExplanation

Module 04

Performance Evaluation Module

Post-quantum algorithms change the physics of your systems: keys grow, signatures grow, latency shifts. The Performance Evaluation Module benchmarks classical algorithms against their post-quantum successors — RSA-2048 vs. ML-KEM-768 for key establishment; ECDSA-P256 vs. ML-DSA-65 and FALCON-512 for signatures; SLH-DSA-128s for hash-based signatures — reporting key/signature sizes and key-generation and signing latencies. Teams see the real-world cost before a single system changes.

  • Classical vs. PQC latency benchmarking
  • Key and signature size comparison
  • Key-generation and signing latency measurement
  • Pre-migration capacity planning data
RSA vs ML-KEMECDSA vs ML-DSA / FALCONLatencySize deltas
Classical baselinePQC candidateBenchmarkReport

Performance evaluation output

Know the cost before you migrate.

Classical vs. post-quantum, measured — key and signature sizes below are the published NIST parameters; AutoPQC benchmarks latency on your own hardware profile.

Key establishment — public key size (bytes)

RSA-2048
256
ML-KEM-768
1,184

Digital signatures — signature size (bytes)

ECDSA-P256
64
FALCON-512
666
ML-DSA-65
3,309
SLH-DSA-128s
7,856

Bars use a square-root scale for readability. AutoPQC benchmarks key-generation and signing latency the same way on your own hardware profile, so the migration cost is measured — not estimated.

What's next — research

Our science moat.

AutoPQC is backed by ongoing doctoral research and peer-reviewed publications. Three directions are moving from the lab into the platform.

R&D 01

Constraint-aware migration sequencing

Planning an estate-wide migration as a constrained optimization problem — respecting dependencies, maintenance windows, and business risk simultaneously.

R&D 02

Calibrated confidence for every recommendation

Statistically trustworthy AI: every recommendation carries a confidence measure you can act on, not just a score.

R&D 03

Governed execution with automatic rollback

Safe, self-correcting deployment — migrations that monitor themselves and reverse automatically when the estate pushes back.

See AutoPQC on your estate.

A demo takes forty-five minutes. The cryptographic inventory it shows you lasts your whole migration.

Request a demo Talk to our team