How it works

AICoven orchestrates multiple models with shared memory. The core ideas: Context Sandwich, Orchestrated routing with fallback, and Roundtable + Reconcile.

1) Context Sandwich

Requests are constructed as a layered bundle:

  1. System prompt with role configuration
  2. Retrieved memories scoped to the current thread/workspace
  3. Recent thread messages (windowed)
  4. Current user message

2) Orchestration with fallback

Each request can target a preferred model but includes provider‑aware fallback paths. If a provider is unavailable or you’ve hit a quota, the router transparently retries on a compatible model you’ve enabled.

3) Roundtable + Reconcile

Multiple roles can respond to the same prompt. A reconcile step compares outputs and synthesizes a single, auditable result you can accept or refine.

Memory approvals

Proposed memories are queued for your review. Approved items become searchable context for future work; rejected items are not retained.