EACI Development Journal · APIs

Why Is a Local AI Fallback Brain Not Cloud Parity?

The local browser engine can preserve basic availability, but it does not pretend to match the cloud model's reasoning or full identity context.

APIs ~8 min issue-tagged
Issue this fixes

Offline and degraded-network support is valuable, especially for status, simple conversation, and known topics. But a compact browser pack has different model capacity, retrieval data, and update timing from the cloud route.

Calling both paths equivalent would create confusing personality and capability shifts. The application instead treats local output as a fallback with a narrower contract.

What was actually going wrong

A single chat interface made it tempting to promise identical behavior from very different engines.

What I tried

What finally worked

Capability checks choose the cloud route when available and reserve the cached browser engine for supported fallback cases.

Why it worked

The narrower promise matches the data and compute actually present on the device.

What I learned

Graceful degradation is honest when it states what was degraded.

Where EACI uses this today

Caelum can use a limited local fallback while Main cloud chat remains the primary brain.

This journal covers real engineering on EACI Companion / The Veil. Companions include Caelum, Chad, Natalia, Atreus, Luna, Roxy, and Cael. Journal articles stay family-safe in content. See Privacy and Ethics.