signal lab · hosted demo
See the substrate work. Before installing anything.
Signal Lab is a hosted dashboard with scripted scenarios and CSV bulk replay — roughly 53,000 events across 1,000 simulated devices, drawn from the six journeys. It is a demo environment: the trust primitive itself runs client-side in production deployments. Signal Lab is the page where the evidence becomes visible.
Four surfaces. Every signal traceable to a journey.
-
·01
Live ticker
Signed Evidence Tokens streaming from the simulated estate in real time — eid, seq, tctx, event, trust basis. Click any token to expand the verification result.
-
·02
Scenario library
Scripted, replayable scenarios drawn from each of the six journeys — ghost transactions, overlay-attack bursts, SIM-swap mid-session, offline-queued retries, restart-interrupted flows, autonomous agent sequences.
-
·03
CSV bulk replay
Roughly 53,000 events across 1,000 simulated devices, replayable at adjustable speed. Sequence-gap surfacing, retry-disambiguation, and reject-reason classification all observable end-to-end.
-
·04
Verifier visibility
Every token is run through the eight-step verifier pipeline; passes and rejects are surfaced with the specific step that decided. No anonymous failures.
Eight reproducible scenarios, one per pattern.
Each scenario runs in seconds, resets cleanly, and ends with a decoded Evidence Token a verifier engineer can read. The full library is wider; these are the eight that operators most often want to see first.
- network
Legitimate retry recovered
Same `tctx`, second `seq`, offline ↔ online transition. Retry recognised and processed once.
- network
Orphan transaction surfaces
Initiated event, no confirmation. Sequence gap visible immediately, not weeks later in reconciliation.
- fraud
Algorithm-confusion attempt
Token presented with weak crypto declaration — refused at verification step ·02, never reaches the risk model.
- fraud
Velocity burst, device-bound
Per-device velocity visible at second-level granularity. Card-testing surfaces in real time, no PII exposed.
- experience
Restart-interrupted payment
Boot-id changes mid-flow. The interruption surfaces; the user is offered follow-up rather than auto-declined.
- integrity
Repeated key rotation
Same device, four rotations in 24 hours. Attributable; reinstall-abuse candidate flagged.
- operations
Cohort outage
100 devices in one province go dark, then flush back. Visible in the ticker before support hears about it.
- autonomous
Agent flow integrity
propose → consent → initiate → confirm — hash-linked. One step omitted; the chain breaks visibly.
Hosted by us. Used by you. No public sandbox.
Signal Lab is a demo environment, not the trust primitive. The substrate itself runs client-side in production deployments — there is no YinkoShield backend in any operator's verification path.
We issue Signal Lab instances on request because each instance is scoped to the journeys, the event volume, and the verifier language an evaluating team actually wants to see. A public sandbox would optimise for footprint over judgement; instances optimise for the conversation that follows.
Typical instance: 60-minute walk-through with the founder; the team gets a dashboard URL valid for two weeks; the team replays the eight default scenarios against their preferred consumption channel (HTTP header, OTel, ISO 8583 DE 48, or SDK direct).
Ask for a Signal Lab walk-through.
We ask for your role and which journeys you want to see — nothing else. We respond within two business days.
Request a Signal Lab instance