CR-Lite — Continuity Receipts Demo

What This Demonstrates

This is a working prototype of the Continuity Receipts (CR) open standard. Every response carries a structured Continuity Receipt showing the epistemic conditions under which it was produced — not just the answer, but the provenance, uncertainty, and review requirements that responsible reliance requires.

When confidence is insufficient for the selected reliance level, the system produces a Failure Receipt instead — turning uncertainty into inspectable infrastructure rather than hiding it in confident prose.

Read the full CR standards proposal →

Select Reliance Level

RC-2 · Research assistance · Minimum aggregate score 0.50 · Human review recommended for consequential decisions

Sample Queries

What are the EU AI Act requirements for high-risk AI systems? What is the legal status of AI-generated works under US copyright law? What does current research show about AI developing consistent behavioral patterns? What are the current federal sentencing guidelines for tax fraud over $1M? What is the precautionary principle in environmental law and how is it applied?

Your Query

PASS RC-2

Response

Continuity Receipt

RC-2 Research Assistance

Aggregate Continuity Score

—

Confidence Dimensions

Source Quality —

Retrieval Coverage —

Internal Consistency —

Temporal Freshness —

Domain Confidence —

Contradictions —

Human Review —

Reliance Level —

This prototype demonstrates the CR standard. A production implementation with full vector retrieval, PostgreSQL audit chains, and cross-system portability is in development. The full technical specification is available at emfoundation.net/pco-standards-schema.html.

CR-Lite — Continuity Receipts

Select Reliance Level

Failure Receipt — Continuity Integrity Insufficient

Continuity Receipt