Built on open standards, verifiable by anyone.

No proprietary lock-in. Industry standards you can trust.

[01] CORE TECHNOLOGIESSTANDARDS, NOT HYPE

Verifiable Credentials

UMP uses W3C Verifiable Credentials 2.0 for signed maintenance events.

  • W3C VC 2.0 data model
  • JSON Web Signature (JWS) for signing
  • did:web for issuer identity
  • Portable, interoperable credentials

Immutable Ledger

Credential hashes anchored to append-only storage for tamper-evidence.

  • AWS QLDB for append-only ledger
  • Sigstore Rekor transparency log (optional, for public auditability)
  • Provable inclusion proofs
  • No speculative tokens or public chains

Asset Identity

Global identifiers that work across systems and scanners.

  • GS1 GIAI standard for asset IDs
  • Digital Link URIs
  • QR/NFC for physical discovery
  • Mapping to existing IDs (VIN, serial, tail)

Telemetry Standards

Native integration with industrial telemetry protocols.

  • ISO 15143-3 / AEMP 2.0 (heavy equipment)
  • OPC UA (industrial) - roadmap
  • NMEA 2000 (marine) - roadmap
  • Custom adapter framework
[02] REGULATORY ALIGNMENTCOMPLIANCE BUILT-IN

Aligned with global standards.

FAA AC 120-78B

Electronic signatures and electronic maintenance records for aviation.

View source →

EU ESPR / Digital Product Passports

Ecodesign for Sustainable Products Regulation establishing DPP framework.

View source →

EU Battery Regulation 2023/1542

Battery passport requirements for EV and industrial batteries by 2027.

View source →

ISO 15143-3 / AEMP 2.0

Fleet data exchange standard for construction and mining equipment.

View source →

USCG Electronic Logbooks

Acceptance of electronic record books for vessels.

[03] SECURITY & PRIVACYTRUST BY DESIGN

Security at every layer.

Keys in KMS

Private keys never leave AWS KMS. Hardware-backed key protection.

S3 Object Lock

Evidence files stored with WORM (Write Once Read Many) protection.

Hash binding

Evidence bound by SHA-256 hash. Any modification is immediately detectable.

Reliable sync

Background task scheduling with exponential backoff ensures credentials sync reliably on any network.

AI never signs.

Humans sign credentials. AI assists with intake and lint, but never has signing authority.

Optional: Public transparency log

For organizations requiring public auditability, UMP supports optional anchoring to Sigstore Rekor — an open-source, publicly verifiable transparency log. Credential hashes become independently auditable by anyone, not just UMP.

[04] EVIDENCE INTEGRITYCANNOT BE FAKED

Tamper-evident evidence binding.

Photos, work orders, and documents are cryptographically bound to credentials. Any modification is automatically detected during verification.

1

Capture & hash

Photo or document captured, SHA-256 hash computed client-side before upload.

2

Store immutably

Evidence uploaded to S3 with Object Lock (WORM mode). Cannot be modified or deleted.

3

Bind to credential

Hash recorded in the signed credential. Evidence URI points to stored file.

4

Verify on demand

During verification, file is re-hashed and compared against credential hash.

Verification results

EVIDENCE_VALID

Hash matches. File is unchanged since credential was signed.

EVIDENCE_TAMPERED

Hash mismatch. File has been modified after credential was signed.

MISSING_EVIDENCE

File not found. Evidence referenced in credential cannot be retrieved.

Evidence cannot be deleted or modified — even by us.

S3 Object Lock in GOVERNANCE mode ensures files are immutable for 7 years by default. Versioning provides a complete audit trail. The "PROTECTED" status shown during verification confirms evidence is in WORM (Write Once Read Many) storage.

[05] AI CAPABILITIESASSISTANCE, NOT AUTHORITY

AI assists. Humans sign.

AI accelerates data entry and improves quality — but never has signing authority. Every credential is signed by a human.

Outputs are traceable

AI summaries cite source events. Confidence indicators show where AI is certain vs. uncertain.

Graceful fallbacks

If AI is unavailable, manual workflows remain fully functional. Core operations never depend on AI.

Privacy-first processing

Document analysis happens in isolated sessions. We don't train on your maintenance data.

Ready to integrate?

Explore the developer documentation or talk to our team.