Skip to content

Axiom Core SDK (v1.0)

Axiom Core is a local semantic transformation SDK. It converts raw input into non-identifying structured context while keeping sensitive data on-device.

This site documents the public SDK. The attested tier depends on a separate native enclave runtime distributed outside this repository. The public SDK includes the simulator path plus verification tools, but does not expose or document private implementation details.

Executive summary

Organizations in regulated domains need cloud AI reasoning without exposing raw data. Axiom Core enables that by transforming sensitive context locally into a semantic structure, enforcing a strict boundary, and (optionally) producing verifiable evidence about the transformation.

The problem Axiom Core solves

Traditional approaches like redaction or anonymization often remove the very signals needed for high-quality reasoning. Pure encryption protects privacy but blocks practical cloud reasoning. Axiom Core preserves meaning while removing identifiers so cloud AI can reason without seeing raw data.

Solution overview

  • Semantic abstraction: extract entities and relations without identifiers
  • Boundary enforcement: raw input never crosses the local-to-cloud boundary
  • Deterministic output: identical inputs produce stable structure
  • Security tiers: standard (software) and attested (TEE + evidence)

Why open source

  • Enables community verification of privacy claims
  • Allows security and compliance teams to audit logic
  • Encourages extensibility across domains
  • Builds trust in boundary enforcement guarantees

Who this documentation is for

  • Developers integrating Axiom Core into production apps
  • Enterprise architects designing private AI systems
  • Security and compliance teams validating boundary guarantees
  • Platform teams evaluating standard vs attested tiers

Roadmap vision (clearly labeled as future work)

  • v1.0 (current): stable local transformation + attested preview
  • v2.x (roadmap): ZKP/zkML verification for transformation correctness
  • v3.x (roadmap): broader ecosystem integrations and verifiable tooling

Market context (selected sources)

The privacy-preserving AI market is projected to grow rapidly over the next decade, driven by regulatory pressure and enterprise demand for secure AI workflows. See sources for estimates and context:

  • Privacy-preserving AI market analysis (Technavio): https://www.technavio.com/report/privacy-preserving-ai-market-industry-analysis
  • Privacy-preserving AI forecast to 2033 (DataIntelo): https://dataintelo.com/report/privacy-preserving-ai-market
  • Private AI market projection (Dimension Market Research): https://dimensionmarketresearch.com/report/private-ai-market/
  • Privacy-enhancing technologies market (Grand View Research): https://www.grandviewresearch.com/industry-analysis/privacy-enhancing-technologies-market-report

Documentation map

  • User Guide: installation, configuration, usage, troubleshooting
  • Developer Guide: architecture, testing, releases
  • API Reference: public types and surfaces
  • Architecture + Security: design and threat model
  • Integration + Examples: verification and deployment patterns

Next: User Guide Overview