Option to automatically hand off to a fallback provider when hitting usage limits

[nathnael-desta]

Problem

Right now, when a thread hits a usage/quota limit (like Claude's 5-hour limit), T3 Code pauses the turn and offers "Resume at reset" or "Snooze until reset".

If you don't want to wait several hours, you currently have to manually change the provider in the model picker (e.g., from Claude to Codex) and send a new message to trigger the provider context handoff.

Idea / Proposal

It would be great to have an option to automatically switch to a configured fallback provider using context handoff when a usage limit is hit, instead of just pausing until the reset window.

• Fallback provider preference: In settings or thread options, let users pick a fallback model (e.g. if Claude Max hits a rate limit, fall back to Codex or Antigravity/Gemini).
• Seamless handoff: When the active provider returns a rate limit / quota error, automatically trigger the existing context handoff mechanism and continue the run on the fallback provider.
• Clear UI indicator: Show an inline indicator or divider in the timeline (similar to the current handoff banner) showing that it switched due to a rate limit.

Since T3 Code already supports both context handoff between providers and auto-resume behaviors, wiring them together would keep long background tasks running without needing manual intervention when a limit hits.

[adamkasper]

i would like to help implement this as an opt in fallback using the existing context handoff

Claude Code Opus 5.5 -> Cursor Opus 5.5 would be ideal when the Claude limit is reached so we can keep using the same model without enabling extra usage in Claude Code when we already pay thousands of dollars for extra usage in Cursor

would you be open to a PR with one fallback provider and model selected by the user