Phylax Verify
Not an opinion. A proof.
Distributed digital authenticity verification. FROST threshold agreement across independent nodes issues a tamper-evident Verifiable Credential bound to your document hash. Any party can verify the result at any time -- without Phylax Verify being involved at all.
The Problem
Every digital object arrives with a question nobody can answer.
Is this email human-written or AI-generated? Was this contract the same document that was signed, or has it been modified since? Is this forensic evidence tamper-free, or was it altered between collection and court? Did this software product actually pass the security audit the vendor claims it passed?
The existing verification tools share a common structural failure. They are centralized. One company operates the model. One company controls the output. That result can be challenged, the system can be compromised, the company can be acquired or pressured. When the verification result matters most -- in a courtroom, in a regulatory submission, in government procurement -- it rests on trusting a single vendor's word.
That is not verification. It is an opinion.
How It Works
Six Steps to a Proof
From submission to an on-chain Verifiable Credential that any party can independently confirm — without calling back to Phylax Verify.
Submit the digital object
A document, email thread, image, video, code repository, audit report, or structured data file is submitted via API, web interface, or SDK. The submission specifies which verification dimensions are required.
Distribution to independent nodes
The object is distributed to N geographically independent Phylax Verify nodes. No single node knows the identity of the submitter. No single node has the ability to alter the verification result.
Independent analysis runs in parallel
Each node independently runs its verification battery against the submitted object. Analyses run in parallel across all participating nodes using separate model architectures.
Threshold agreement computed
The coordinator aggregates per-node results. For a result to be issued, K of N nodes must agree within defined tolerance thresholds on each verified dimension. Where fewer than K nodes agree, the dimension is marked inconclusive rather than producing a false confidence score.
Verifiable Credential issued
A W3C Verifiable Credential is issued containing the verified object hash (SHA-256), the verification timestamp, each dimension result with the number of agreeing nodes, and a FROST threshold signature from the node network. The credential is anchored on-chain.
Third-party verification -- anytime, without us
Any party with the credential ID and the document can independently verify the result by presenting both to any Phylax Verify node. The node checks the on-chain credential, recomputes the hash, and confirms whether the document matches the verified original. Phylax Verify does not need to be involved.
Verification Dimensions
What Each Node Analyzes
Each verification request specifies which dimensions to run. Nodes analyze independently. Threshold agreement is required for each dimension to appear in the credential.
AI Generation Probability
Does the content exhibit structural, linguistic, or statistical patterns consistent with current large language models, image generators, or video synthesis models? Per-node detection models update continuously from the cross-node intelligence feed. Threshold agreement across diverse independent model architectures is resistant to adversarial generation techniques designed to evade a single detector.
Tamper Detection
Does the object match a pattern consistent with unmodified content? Nodes check internal structure, metadata chain, hash history, and content signature. Modification artifacts flagged include revision history erasure, metadata timestamp inconsistencies, formatting artifacts from format conversion, and invisible character injection.
Provenance Dating
When was this content most likely created? Each node produces an estimated creation date range from internal metadata, linguistic and stylistic signals, referenced content timestamps, and file format generation markers. Threshold agreement on the range produces a confident provenance date.
Process Attestation
For structured compliance submissions -- audit reports, security assessments, procurement documents, financial filings -- did the document originate from the stated process? Process attestation checks whether the document structure, control coverage, and terminology are internally consistent with a genuine assessment outcome.
Identity Binding
If the submitter presents a PhylaxOne identity (DID), the verification credential can bind the object to that identity with a cryptographic proof. Not required for all use cases -- available as an optional dimension for chain-of-custody and regulatory submissions where submitter attribution matters.
Per-Verification Pricing
$2 -- $25
AI detection only starts at $2. Full battery (all five dimensions) is $25 per verification.
Use Cases
Where a Proof Matters
Legal and Digital Forensics
Digital forensic evidence has a chain-of-custody problem existing tools do not solve. An investigator collects a disk image, memory dump, or communication record. Between collection and court, that evidence passes through multiple hands and systems. Current chain-of-custody documentation relies on procedural attestation -- a signature on a form, not cryptographic proof. Phylax Verify issues a credential at the moment of collection binding the hash of the evidence to a threshold-signed proof. Any future presentation of the evidence can be checked against that hash. If it matches, the evidence is the same object. If it does not, the modification is detectable and attributable.
Compliance Submissions and Audit Reports
CMMC assessors, FedRAMP reviewers, and regulatory bodies receive documents every day claiming to represent genuine assessments. Currently there is no mechanism to verify that a compliance report reflects an actual assessment rather than a pre-filled template. Phylax Verify's process attestation dimension checks whether a submission is internally consistent with a genuine assessment outcome. An AXIOsecurity-verified CMMC System Security Plan tells the C3PAO assessor: this document was verified by K of N independent nodes as consistent with a genuine NIST 800-171 assessment. The credential is on-chain. The assessor can verify it independently.
Contract and Document Authenticity
Every enterprise legal team handles contracts, NDAs, amendments, and agreements where the question of whether a document has been modified since signing is answerable only by comparing to a saved copy -- if one exists. Phylax Verify issues a credential at signing time binding the document hash to the credential. Any future dispute about whether the contract has been modified since signing is resolved by checking the hash, not by litigation over saved versions. For M&A due diligence, IP assignments, real estate transactions, and government contracting, document authenticity becomes a mathematical fact rather than a legal argument.
AI-Generated Content Detection at Scale
Media organizations, academic institutions, government communications offices, and enterprise legal teams all face the same question: is this content human-authored? Current tools answer with a percentage score from a single model -- a score that can be challenged, gamed, or simply wrong. Phylax Verify answers with a threshold-agreed result from distributed independent nodes, each running a different model architecture. A document that scores as AI-generated across K of N independent nodes, each running a different model, is a meaningfully stronger result than any single model score. Threshold agreement is resistant to adversarial generation techniques designed to evade a specific detector.
Comparison
Existing Tools vs. Phylax Verify
Existing Tools
Single company, single model
Phylax Verify
Threshold agreement: K-of-N independent nodes
Existing Tools
One company's output -- contestable in court
Phylax Verify
Mathematically provable threshold credential -- on-chain
Existing Tools
No tamper detection after verification
Phylax Verify
Hash-bound credential detectable by any party forever
Existing Tools
Requires calling back to the vendor to verify
Phylax Verify
Public on-chain credential, verifiable without vendor
Existing Tools
One model to game with adversarial content
Phylax Verify
K diverse models in threshold -- gaming one changes nothing
Existing Tools
No quantum resistance
Phylax Verify
ML-DSA (NIST FIPS 204) signatures on all credentials
Existing Tools
AI content detection only
Phylax Verify
AI detection, tamper, provenance, process attestation, identity binding
Existing Tools
Not designed for court or regulatory use
Phylax Verify
Cryptographic chain of custody, hash-bound, admissibility-ready
AXIOsecurity
The Engine Behind the Certification Mark
When a product, document, or process earns the AXIOsecurity Certified mark on AxioTerra, that mark is a Verifiable Credential issued by Phylax Verify nodes in threshold agreement. Not a badge from a single auditor. Not a self-attested claim. A cryptographically provable result from an independent distributed network.
Every verification event also generates behavioral intelligence that feeds back into the AXIOsecurity certification criteria. What attack patterns are currently active. What structural vulnerabilities are appearing in real DFIR investigations. What new AI generation signatures are showing up in real submissions. The criteria update from real-world adversarial data, not committee publication schedules.
A product certified against AXIOsecurity criteria today is tested against what attackers are doing today -- not what a committee documented two years ago.
Pricing
Simple Subscription Tiers
All tiers include full API access, all verification dimensions, and the on-chain anchored Verifiable Credential output. AXIOsecurity certification assessments are available separately.
Developer
$299 / month
5,000 verifications. Full API access. All verification dimensions.
Business
$1,500 / month
50,000 verifications. SLA. Priority node access.
Enterprise
$6,000 / month
Unlimited verifications. Dedicated node allocation. White-label credential branding.
Legal and DFIR
$3,500 / month
Per law firm or DFIR firm. Unlimited chain-of-custody verifications.
AXIOsecurity Certification
$5,000 -- $15,000 per assessment
Annual renewal at $2,500 -- $7,500. MSSP white-label licensing available.
Ready to replace the opinion?
All briefings conducted under NDA. Bring your use case -- evidence chain of custody, compliance submission verification, contract authenticity, or AI content detection at scale. We will walk through the verification workflow, the credential format, and the integration path.