17 / Integrations

Reference integrations will appear only after verification.

This register separates an application, a selected proposal, a working pilot and an accepted reference integration. No third-party integration is counted before its public evidence passes the challenge rubric.

Verified integrations0Open reference slots3Independent auditPending
Current evidence:The Receipt API beta and Proof Engine work publicly. Third-party integrations, customers and production settlement are not currently claimed.Apply to build

00 / Reference tracks

Three runnable starting points, clearly labelled first-party.

These adapters reduce integration work while preserving the distinction between a PactVerity example and a third-party implementation.

JavaScriptRunnable

Agent receipt adapter

Create and verify a receipt from a Node.js agent or API workflow using the zero-dependency preview client.

Open acceptance guide
PythonRunnable

Automation receipt adapter

Connect a Python workflow to the public Receipt API and preserve a reproducible verification result.

Open acceptance guide
EdgeBlueprint

Minimal receipt ledger

Verify first, then retain a privacy-conscious audit index without presenting storage as evidence truth.

Open architecture guide

Download runnable SDK previews and examples

01 / Evidence gates

Four stages, four different claims.

Stage 01Private

Application received

A developer has proposed work. This is not an integration or endorsement.

Stage 02Selected

Proposal accepted

Scope and acceptance criteria are agreed. No working result is claimed.

Stage 03Pilot

Demonstration running

A build exists, but reproducibility and security review may remain incomplete.

Stage 04Verified

Reference integration

Public code, receipt, demo and limitations have passed the published rubric.

02 / Case-study record

Every accepted integration will disclose the same facts.

  • Developer, project and public repository.
  • Real service or test environment and the problem being measured.
  • Exact PactVerity API surface used and a sample receipt.
  • Reproduction steps, result date and known limitations.
  • Challenge score, reviewer identity and material relationships.
  • What the case study does not prove: customers, revenue, audit assurance or token value.

Open the public case-study template