[Feature]: OpenHands-managed provider fallback for Codex ACP conversations
Maintainers usually reply within 1 day
Nobody has claimed this yet.
Assessment
- Difficulty
- 5/5
- Estimated time
- Over a week
- Newbie friendliness
- 22/100
Research direction
This is a design-level request: no file, class or test is named, and the reporter explicitly leaves the scope as "Unknown". Start by reading how Codex ACP conversations are configured and launched in the Agent Server, and where the existing native OpenHands LLM fallback is wired, to judge whether provider switching can be layered on without touching Codex itself. Done means the four acceptance criteria hold: opt-in bounded failover, an explicit report of the provider/model change, workspace preservation with a clear stop when continuity fails, and no credential leakage or tool-action replay. A maintainer should decide feasibility and shape before anyone starts coding.
Written by the indexing model from the issue text.
Description
Is there an existing feature request for this?
- I have searched existing issues and feature requests, and this is not a duplicate.
Problem or Use Case
Hi team! Would you consider adding opt-in provider fallback managed by OpenHands for Codex-backed ACP conversations?
My use case is a ChatGPT-backed Codex primary, with a separately authenticated Kimi Coding backup when the primary hits its quota or has a temporary outage. I'd like to keep Codex as the execution engine rather than switch to a different agent.
Kimi documents custom-provider support through Codex, so this is a request for a general orchestration capability, not a new provider-specific integration.
Desired Behavior
Let users configure a primary and backup provider/model, their credential references, and bounded failover conditions at the OpenHands layer, without requiring Codex itself to implement automatic failover.
The important part is a safe, visible handover. I'm not assuming native Codex session state can simply transfer between providers. If a fresh session or reconstructed context is needed, please make that clear and stop if safe continuation isn't possible.
Acceptance Criteria
- Failover is opt-in, bounded, and triggered by configured conditions such as quota exhaustion or temporary provider failures.
- Codex remains the execution engine, and the actual provider/model change is reported explicitly.
- The workspace is preserved, with safe context recovery or a clear stop when continuity cannot be established.
- Credentials remain protected, completed tool actions are not replayed, and permission or policy denials are not bypassed.
Alternatives Considered
Native OpenHands LLM fallback applies to a different inference path. openai/codex#32341 requests failover inside Codex; this request is for OpenHands-owned orchestration that can be considered independently.
Automatically routing work to competing providers may not align with OpenAI's commercial incentives, so they may be less interested in prioritising that request. An OpenHands-managed implementation would avoid that dependency and keep fallback decisions provider-neutral.
Priority / Severity
Medium - Would improve experience
Estimated Scope
Unknown - Not sure about the technical complexity
Feature Area
- Agent API / Core functionality
- Agent Server
- Configuration / Settings
Additional Context
If there's already a supported way to do this for Codex ACP conversations, a pointer to its configuration and continuity guarantees would be very welcome. Thanks for taking a look!
- Dominant language
- Python
- Stars
- 1.4k
- Forks
- 797
- Avg merge
- 1d 17h
- Merged PRs (30d)
- 354
Getting set up
- No Dockerfile or Docker Compose file
- Has a pull request template
- Read the contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from OpenHands/software-agent-sdk
-
[Bug]: `list_directory` reports a complete listing as truncated when a directory has exactly 500 entriesPossibly taken @balabommablock-cpu claimed this 29 days ago. Openbug priority:low ready-for-dev
Difficulty 2/5 1-3 hours Newbie friendliness 87/100
OpenHands/software-agent-sdk#6133 ·
Maintainers usually reply within 1 day
-
[Bug]: Command hook that exits 1 is treated as blocking when its stdout is a JSON `deny` or `continue: false`Possibly taken A pull request linked to this issue is open or already merged. Openbug priority:low ready-for-dev
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
OpenHands/software-agent-sdk#6116 · 3 comments ·
Maintainers usually reply within 1 day
-
[Bug]: TypeScript GET-with-body requests remain pending after a truncated responsePossibly taken @BlueRaddish claimed this 34 days ago. Openbug priority:low ready-for-dev
Difficulty 2/5 1-3 hours Newbie friendliness 85/100
OpenHands/software-agent-sdk#6059 ·
Maintainers usually reply within 1 day
-
[Bug]: Document whether `prompt_tokens` includes `cache_read_tokens` in accumulated token usagePossibly taken @Jaxon1216 claimed this 36 days ago. Openpriority:low ready-for-dev
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
OpenHands/software-agent-sdk#5975 ·
Maintainers usually reply within 1 day
-
priority:low
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
OpenHands/software-agent-sdk#5708 · 1 comment ·
Maintainers usually reply within 1 day
All issues in OpenHands/software-agent-sdk
Similar issues
-
[Bug]: reme CLI crashes when the running server used a relative config pathPossibly taken @nurlanmnn claimed this today. Open
Difficulty 2/5 1-3 hours Newbie friendliness 88/100
agentscope-ai/ReMe#770 ·
Maintainers usually reply within 1 day
-
Pi: paused model calls are reported as "not signed in" when the model's provider comes from an extensionPossibly taken @balcsida claimed this today. Open
Difficulty 2/5 1-3 hours Newbie friendliness 73/100
microsoft/TokenEconomizer#288 ·
Maintainers usually reply within 1 day
-
Difficulty 1/5 Under an hour Newbie friendliness 67/100
sickn33/agentic-awesome-skills#2437 ·
Maintainers usually reply within 1 day
-
site-bug triage
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
ethereum/consensus-specs#5846 ·
Maintainers usually reply within 1 day