THE NEUTRAL TRUST LAYER

The authenticity authority
for digital evidence.

CanonProof is the neutral trust layer for the internet — cryptographically certifying files and exposing explainable forensic reasons behind every result.
Use CanonProof to strengthen evidence before it’s relied upon: issue a signed certificate, attach a public verification link, and cite clear, human-readable reasons in your report.
  • Neutral & standards-driven — built for scrutiny, not “truth claims”
  • Legal-grade certificates — ES256/JWS signed payloads with stable IDs
  • Explainable forensics — indicators translated into reasons you can cite
  • Open verification — public verify URL + JWKS for independent checks
Public verify How it works Create account
ES256-signed certificates • SHA-256 hashing • Public verification link
Latest certificate (example)
Likely Authentic
Certificate summary
Label: LIKELY_AUTHENTIC
Score: 92
Confidence: 0.94
Media type: image/jpeg
Certificate ID: CP1-9QF2-7K2M-1A0X
SHA-256: e5c1…a4bd
Verify URL: /certificates/…
Public verification: Anyone can validate the signed result (no account required)
Signed payload: ES256/JWS + public JWKS for independent validation
CanonProof certifies integrity and authenticity signals — it does not decide “truth.” AI indicators (when available) are advisory only and never the sole deciding factor.

Built for scrutiny

CanonProof creates a defensible verification record for images, video, audio, and PDFs. Each result is cryptographically bound to the file, explained in plain English, and shareable via a public verification link.

Shield check icon
Certificate

Legal-grade certificates

Stable certificate IDs, SHA-256 fingerprints, and ES256/JWS signatures — designed to be attached to case files, reports, and records.

  • Cryptographically tamper-evident
  • Timestamped and repeatable
  • Built for audit trails
List search icon
Explainability

Explainable reasons

Indicators are translated into human-readable reasons (metadata plausibility, compression artefacts, tamper signals, and more) so conclusions are easier to defend.

  • Clear “why” behind the label
  • Evidence-style wording (not hype)
  • AI signals are advisory only
Link icon
Verification

Open verification

Every certificate includes a public verification URL and signed payload so third parties can validate results without logging in.

  • No account required to check
  • Same result for everyone
  • Designed for disclosure

How CanonProof certifies authenticity

CanonProof creates a repeatable, defensible verification record. Every step is deterministic, explainable, and cryptographically bound to the file.

  • 1 Hash the file
    The file is fingerprinted using SHA-256 to create a stable, tamper-evident identifier.
  • 2 Extract indicators
    Metadata plausibility, compression artefacts, tamper signals, and format-specific checks are evaluated.
  • 3 Score & label
    Indicators are weighted to produce a consistent label and confidence score.
  • 4 Sign & publish
    A certificate is issued, ES256-signed, and made available via a public verification link.
View the full methodology →
Digital forensics and verification workflow diagram

Explainable reasons you can cite

CanonProof doesn’t just label files — it explains why. Forensic indicators are translated into clear, human-readable reasoning that maps directly to the final outcome. This gives you language you can actually cite when evidence is questioned.

Each result includes both the underlying indicators and a plain-English interpretation of what they mean. That distinction matters in disputes: one shows what was measured, the other explains how it was interpreted.

Outcomes are consistent and limited to four clear states: AUTHENTIC, LIKELY_AUTHENTIC, INDETERMINATE, or MANIPULATED. AI analysis (when available) is treated as an advisory signal only — never the sole deciding factor.

Try public verify →
Public verify excerpt · Indicators & Reasoning
LIKELY_AUTHENTIC
Score: 92
Confidence: 0.94
Media: image/jpeg
Certificate: CP1-DR5JVSF…
Indicators Reasoning
Computed indicators (example)
img.exif.make_model present
img.thumb_mismatch false
img.ela.p95 0.016
img.jpeg.quant_table consistent
ai.image.ai_prob 0.11 (advisory)
Indicators Reasoning
Plain-English reasoning (example)
• Camera make/model metadata is present and internally consistent.
• Embedded thumbnail matches the main image, reducing likelihood of post-capture recomposition.
• Low ELA residuals are consistent with a single export rather than repeated edit-save cycles.
• JPEG structure appears consistent with a standard camera pipeline.
• AI signal indicates low synthetic probability and is treated as advisory only.
Designed for disputes: indicators show what was measured, and reasoning shows how CanonProof interprets it — wording that can be copied into reports and independently verified via the public verification page.
Who relies on CanonProof

Verification for high-stakes workflows

Bring trust back to digital evidence

Issue your first CanonProof certificate in minutes — or embed verification into your platform with the API.

Create account Public verify See pricing

Public verification requires no account • Plans start at £499/mo

Pricing built for high-stakes evidence

CanonProof is priced for professional workflows where evidence must stand up to scrutiny. Every plan includes cryptographic certification, a public verification link, and explainable reasons you can cite.

Monthly
Yearly
Public verification requires no account • Built for scrutiny • CanonProof certifies integrity & authenticity signals — not “truth”

Frequently asked questions

CanonProof is built for high-stakes evidence workflows. These are the questions most teams ask before adopting it.

CanonProof certifies a file’s integrity and authenticity signals. It hashes the file (SHA-256), evaluates format-specific forensic indicators, and issues a signed certificate with a clear outcome and confidence. It does not determine “truth”, intent, or context — it provides a defensible verification record.

Certificates are cryptographically signed (ES256/JWS) and bound to the file’s SHA-256 fingerprint. If the file changes, the fingerprint changes. If the certificate is altered, the signature validation fails. Third parties can verify the signature via the public verification page (and public keys where applicable).

Outcomes are deliberately limited to four clear states so teams can standardise decisions:

AUTHENTIC / LIKELY_AUTHENTIC indicate no material manipulation signals were detected (with differing confidence).
MANIPULATED indicates indicators consistent with alteration are present.
INDETERMINATE means the file cannot be assessed confidently (e.g., insufficient signal, format limits, or conflicting indicators).

Every outcome includes explainable reasons so it’s clear what drove the result.

AI analysis may be included as an advisory signal when available, but it is never the sole deciding factor. CanonProof’s outcome is driven by explainable forensic indicators and a transparent scoring approach. AI signals are presented separately and treated as supporting context, not proof.

No. Each certificate includes a public verification link so third parties can validate the signed result without logging in. This is designed for disclosure — clients, auditors, legal teams, and other stakeholders can view the same outcome and verify the signature independently.

CanonProof is built for professional evidence workflows where retention and auditability matter. Certificates store verification outputs (fingerprints, outcome, confidence, indicators, and reasoning). Retention depends on your plan and settings. For sensitive workflows, you can control what is shared publicly (public verify exposes the verification record needed for validation, not your case notes).

Bring trust back to digital evidence

Issue your first CanonProof certificate in minutes — or embed verification into your platform with the API.

Create account Public verify See pricing

Public verification requires no account • Plans start at £499/mo