The live prototype compares two independent GPT-5.6 responses to the same business question:
• a context-free Baseline response, and • a NESTOR response using only the user-enabled signals from a synthetic Human Context Passport.
The current release includes:
• live GPT-5.6 Responses API integration, • transparent Human Context Passport controls, • Context Used attribution, • Personalisation Delta, • deterministic evaluation panels, • strict Structured Outputs and Zod validation, • privacy-preserving diagnostics, • and safe static fallback behaviour.
After reliability hardening, the production deployment completed five consecutive successful live comparisons.
Live demo: https://nestor-day-one-agent.vercel.app
Log in or sign up for Devpost to join the conversation.