Agents don't work when there are new model releases!

Hi Base44 team,

When Opus 4.8 was released yesterday, agents I'd configured to use Opus 4.7 stopped working — the API call to 4.7 failed, and the agent had no way to recover. The agent just broke until I manually updated the model.

I'd like to request a fallback model feature in the agent creation flow. Specifically:

1. Let me set a primary model (e.g., Opus 4.8) and one or more fallback models (e.g., Opus 4.7, Sonnet 4.6) in priority order.

2. If the primary model call fails — deprecated model, API error, rate limit, capacity — the agent should automatically try the next model in the fallback chain instead of erroring out.

3. Ideally, surface a notification/log when a fallback was used, so I can see it happened and update the primary later.

Why this matters:

• Model releases shouldn't break production agents.

• Deprecations are predictable but the timing isn't always — fallbacks make agents resilient.

• This is standard practice in other LLM orchestration tools (LangChain, LiteLLM, etc.) and would put Base44 at parity.

Thanks!