OpenCode Model Router: per-agent model fallback chains with a local web UI

Long sessions tend to end the same way: the primary model starts returning rate-limit errors, or the provider goes quiet mid-task. The work is half done, the context is fully loaded, and the fix is either to wait or to switch models by hand and hope continuity survives. If you run more than one provider — OpenRouter, DeepSeek, and others — you already have somewhere to fail over to. What is usually missing is a way to declare "this agent tries A, then B, then C" and have that applied automatically when A stops answering. Waiting out a rate limit is fine once; doing it repeatedly, across agents