Synthetic Users architecture

Current public deployment is a static beta intake site. Target product architecture applies MiroShark-style simulation and Aeon-style autonomy to pre-launch reaction testing.

current public deployment target product runtime visitor founder / PM / team static site single-file landing page marketing + beta intake ai.syntheticuser.workers.dev optional wallet window.ethereum Rabby / MetaMask localStorage context beta form payload xHandle, contact, URL summary, attack surface JSON POST external inbox webhook.site endpoint stores beta requests no local backend in bundle current truth public artifact today = landing + wallet context + external beta intake launch surface landing page / thread onboarding / pricing / spec intake + normalization fetch / clean / classify extract claims + structure persona selection choose synthetic set patience + objections MiroShark-style simulation layer skeptic trust / proof attack buyer ROI / payoff scan operator friction scan confused first-time visitor reaction aggregation cluster repeated objections rank leaks + confusion decision memo what landed what killed trust / rewrite next rewrite + rerun loop fix hero / proof / ordering rerun before launch Aeon-style orchestration layer repeatable run loop run -> summarize -> act -> rerun rerun Legend Frontend / intake Runtime / output Simulation personas Security / wallet External / infra

What exists today

  • • Public site is a static HTML deploy
  • • Optional wallet context via injected wallet
  • • Beta form posts JSON to an external webhook inbox
  • • No local backend is bundled into the public artifact

MiroShark contribution

  • • One input, many reacting synthetic personas
  • • Disagreement is signal
  • • Pressure beats fake validation
  • • Good for first-reaction simulation

Aeon contribution

  • • Repeatable autonomous run loop
  • • Run, summarize, act, rerun
  • • Artifact-first output instead of chat sludge
  • • Applied here to launches and onboarding