[FEATURE]: Support multiple API keys per provider with fallback logic

[yogeek]

Describe the enhancement you want to request

as #1326 has been closed and it seems the feature has gained some interest, here is another issue as suggested

Problem

Under heavy usage, multiple Opencode instances hit rate limits, causing failures and interrupting sessions.

Proposed Solution

Allow users to configure multiple API keys for each LLM provider and automatically fallback to the next key when the current one returns a rate limit error (e.g. HTTP 429) or other failures.

On each request:

• Try with the primary key.
• If rate‑limited or failed, automatically retry with the next key.
• Continue until a key succeeds or all fail.
• Provide optional logs/metrics showing fallback usage (which key, error type).

[github-actions]

This issue doesn't fully meet our contributing guidelines.

What needs to be fixed:

• The description field lacks meaningful content. A feature request should explain the problem being solved, the proposed solution, and relevant context — not just reference a closed issue. Please describe what "multiple API keys per provider with fallback logic" should look like, why it's needed, and how it should work.

Please edit this issue to address the above within 2 hours, or it will be automatically closed.

---

This issue might be a duplicate of existing issues. Please check:

• Support multiple API keys per provider with fallback logic #1326: Same feature request (multiple API keys per provider with fallback logic) — this issue explicitly references it as its basis

If you believe this was flagged incorrectly, please let a maintainer know.

[github-actions]

Thanks for updating the issue! It now meets our contributing guidelines. Your feature request is clear and well-described.

[LeoLP1]

Is there any news about this feature?