# InterAI Risk Oracle

InterAI Risk Oracle is an independent pre-execution decision layer for consequential agent actions. Before an agent executes, InterAI verifies the proposed action, applies authoritative policy, and returns allow, review_required, or block with signed trust receipt metadata. InterAI does not execute the external action.

## Start Here

- Demo: https://api.interailabs.dev/demo
- Action Boundary Lab: https://api.interailabs.dev/lab
- Pilot application: https://api.interailabs.dev/pilot
- Pricing: https://api.interailabs.dev/pricing
- OpenAPI: https://api.interailabs.dev/.well-known/openapi.json
- Service descriptor: https://api.interailabs.dev/.well-known/ai-service.json
- Discovery bundle: https://api.interailabs.dev/.well-known/discovery-bundle.json
- Agent card: https://api.interailabs.dev/.well-known/agent.json
- Autonomous adoption: https://api.interailabs.dev/.well-known/autonomous-adoption.json

## Safety Notes

- The hosted demo uses a limited demo_trial key for one safe read-only sandbox flow.
- Demo traffic does not move funds, confirm topups, execute external tools, fetch arbitrary URLs, or bypass payment for non-trial traffic.
- External evidence is parsed, verified, bound, and recorded without changing signals, score, semantic_judge, policy, or the final decision in Stage 2.