Live request path

RC + RS + AETHER + URL-DB + SARCH

The current project architecture has one serving model: RC owns projects and public routing, RS handles aggressive cache and continuity, RC Storage exposes project databases, AETHER/URL-DB persist data, and SARCH archives history outside the live request path.

A project request enters the RC front door. RS may answer immediately; otherwise RC resolves or wakes the execution and serves static AETHER content or proxies to the private runtime.

Project browser ↓ RC front door ↓ RS cache / visitor continuity ├─ HIT → response now + optional background wake └─ MISS → RC execution → response

RC Storage uses URL-DB for logical table/row documents and AETHER for typed references, content hashes, cache records, policies, snapshots and binary chunks.

RS can hibernate inactive runtimes while preserving safe durable data and per-visitor state. A new request can be cache-first and wake the runtime separately.

SARCH preserves selected historical page evidence under strict rate limits. It does not run as a prerequisite for serving current pages.