INTEGYRS
04 / Evidence

Show the result.
Keep the context.

A dated snapshot of our experimental software and isolated private-node work. Results are bounded by the setup, the procedure and what the test actually observed.

REVIEWED 18 OCTOBER 2026INTERNAL ENGINEERING RECORDSNO INDEPENDENT AUDIT CLAIMED
195offline tests
47private-node checks
2forced race orders

Counts describe different suites and are not added into a single security score. These are recorded checkpoint results; building this website does not rerun the product tests.

Claim register

What was tested.
What it means.

These summaries keep an observation and its limitation together. Only explicitly approved source material is made public. Private source code and raw run logs remain outside this website.

E01 / 18 October 2026Software alpha

An executable ownership lifecycle

Offline software test environment

The recorded checkpoint reports 195 passing offline tests covering cryptographic, lifecycle and failure behavior.

PROCEDURE

Run the versioned offline test suite; inspect expected successes and deliberate refusal cases.

LIMITATION

Software nodes and endpoints do not establish secure hardware, production cryptography or an independent audit.

Source: Internal engineering test records; not publicly released.

E02 / 18 October 2026Private-node experiment

Transactions on an isolated node

Isolated BSV regtest; no public peers

A separate private-node run reports 47 main checks for bounded mint, transfer, conflict, rollback and recovery scenarios.

PROCEDURE

Exercise candidate transactions against a private node and compare the resulting ownership and access observations.

LIMITATION

This is regtest evidence. Public testnet, mainnet settlement and production availability are not demonstrated.

Source: Internal private-node run records; not publicly released.

E03 / 18 October 2026Bounded software experiment

Both sides of the freeze race

Forced software race orders with private-node state

Freeze-first denies the seller’s new authorization and retained, unconsumed capability use. When an authorization receipt commits first, that already-authorized delivery may finish after freeze.

PROCEDURE

Force freeze-first and receipt-first orderings; check receipts and access outcomes against the ordering rule.

LIMITATION

This is a software authorization boundary. It does not erase copied plaintext or prove continuous protected rendering.

Source: Internal authorization-race records; not publicly released.

Reading the evidence

Different tests answer
different questions.

Functional tests check expected software behavior. Adversarial experiments probe a specified attack or failure. A private-node run checks bounded transaction and access scenarios in isolation.

Independent cryptographic review, production reliability, public-network settlement and actual protected-device operation require their own evidence. One category does not stand in for another.

Not yet established

The remaining assurance work.

  • A real protected reader with verified attestation and display control
  • Public testnet or mainnet end-to-end settlement
  • Independently deployed access nodes and production key custody
  • Completed independent cryptographic and implementation review
  • Measured availability, support commitments or a commercial service
Source provenance

An evidence trail that can be reviewed.

The current engineering results are summarised from a reviewed version. Underlying private records are not publicly downloadable; any later release should preserve a stable version, redaction boundaries and reproducible procedure. Public standards and patent records are linked from the technical chapters.

New results should update this register before headline wording changes. A failed or incomplete gate stays visible.

How to read the expanded technical library

Different sources.
Different claims.

The depth of an explanation does not change the strength of its evidence. Each chapter identifies which statements describe the alpha, public reference mechanisms or a proposed architecture.

01 / INTERNAL SOFTWARE RECORD

Observed, bounded behavior.

The protocol chapter explains the reviewed experimental software and isolated-node record. Counts are not a fresh independent rerun, production benchmark or security certification.

Read the current protocol →
02 / PUBLIC PRIMARY SOURCES

Mechanisms and publications.

Teranode documentation, firmware/attestation references and patent publications establish their stated facts. They do not certify an INTEGYRS implementation or confer patent rights on it.

Read the publication register →
03 / FUTURE DESIGN

Requirements to demonstrate.

Competing threshold providers, compulsory hardware, no full-work plaintext object and protected output are architecture targets. The test plan is explicit precisely because those properties are not yet established.

Read the acceptance tests →

The next claim needs
the next piece of evidence.

Follow the roadmap