openclaw - ✅(Solved) Fix Docs/UX: clarify ChatGPT Plus vs Codex OAuth vs API-key routing for GPT-5.4 (and document limit differences) [1 pull requests, 1 comments, 2 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openclaw/openclaw#61666Fetched 2026-04-08 02:56:13
View on GitHub
Comments
1
Participants
2
Timeline
1
Reactions
0
Timeline (top)
commented ×1

OpenClaw currently makes it hard to understand how ChatGPT Plus / ChatGPT OAuth, Codex OAuth, and OpenAI API key relate when using GPT-5.4.

Two user-facing problems keep causing confusion:

  1. Users cannot tell whether ChatGPT Plus entitlement can be used directly for openai/gpt-5.4, or only indirectly via openai-codex/gpt-5.4.
  2. The usable quota/limits through openai-codex/gpt-5.4 appear significantly tighter than the ChatGPT web experience, but this difference is not clearly explained in OpenClaw UX/docs.

Root Cause

OpenClaw currently makes it hard to understand how ChatGPT Plus / ChatGPT OAuth, Codex OAuth, and OpenAI API key relate when using GPT-5.4.

Two user-facing problems keep causing confusion:

  1. Users cannot tell whether ChatGPT Plus entitlement can be used directly for openai/gpt-5.4, or only indirectly via openai-codex/gpt-5.4.
  2. The usable quota/limits through openai-codex/gpt-5.4 appear significantly tighter than the ChatGPT web experience, but this difference is not clearly explained in OpenClaw UX/docs.

PR fix notes

PR #63443: Clarify OpenAI auth setup hint for separate Codex OAuth

Description (problem / solution / changelog)

Summary

Describe the problem and fix in 2–5 bullets:

  • Problem: The OpenAI auth/setup hint still implied a combined OpenAI/Codex auth path even though openai (API key) and openai-codex (OAuth) are now separate provider routes.
  • Why it matters: This creates onboarding/configure confusion and can leave users thinking the standard OpenAI setup also covers Codex OAuth.
  • What changed: Updated the OpenAI group hint text to say API key only (use separate OpenAI Codex for OAuth) in both the plugin manifest and provider auth choice definition. Happy to reword as needed
  • What did NOT change (scope boundary): No auth logic, provider routing, token handling, or model resolution behavior changed. Its just a string change

Change Type (select all)

  • Bug fix
  • Feature
  • Refactor required for the fix
  • Docs
  • Security hardening
  • Chore/infra

Scope (select all touched areas)

  • Gateway / orchestration
  • Skills / tool execution
  • Auth / tokens
  • Memory / storage
  • Integrations
  • API / contracts
  • UI / DX
  • CI/CD / infra

Linked Issue/PR

  • Closes #
  • Related #30533
  • Related #61666
  • This PR fixes a bug or regression

Root Cause (if applicable)

For bug fixes or regressions, explain why this happened, not just what changed. Otherwise write N/A. If the cause is unclear, write Unknown.

  • Root cause: UX copy drift. The setup hint kept older combined wording after openai and openai-codex became separate auth/provider paths.
  • Missing detection / guardrail: No test or copy audit was asserting that onboarding/configure text clearly distinguishes OpenAI API-key auth from Codex OAuth.
  • Contributing context (if known): There was prior issue history around this confusion, but the current hint on main was still stale.

Regression Test Plan (if applicable)

For bug fixes or regressions, name the smallest reliable test coverage that should catch this. Otherwise write N/A.

  • Coverage level that should have caught this:
  • Unit test
  • Seam / integration test
  • End-to-end test
  • Existing coverage already sufficient
  • Target test or file: auth-choice / onboarding output coverage for OpenAI provider hints
  • Scenario the test should lock in: OpenAI API-key setup text should clearly say it is API-key only and point users to separate OpenAI Codex OAuth when relevant.
  • Why this is the smallest reliable guardrail: The bug is in assembled UX copy, not auth runtime behavior.
  • Existing test that already covers this (if any): Existing checks/lint passed for touched files, but I did not find a test explicitly locking this wording in.
  • If no new test is added, why not: This PR is intentionally tiny and copy-only; I kept scope limited to correcting the stale hint.

User-visible / Behavior Changes

List user-visible changes (including defaults/config). If none, write None.

  • OpenAI setup/auth hint now reads: API key only (use separate OpenAI Codex for OAuth)

Diagram (if applicable)

For UI changes or non-trivial logic flows, include a small ASCII diagram reviewers can scan quickly. Otherwise write N/A.

Before: [user selects OpenAI setup] -> [hint suggests combined OpenAI/Codex auth path]

After: [user selects OpenAI setup] -> [hint says API key only] -> [user is directed to separate Codex OAuth path if needed]

Security Impact (required)

  • New permissions/capabilities? (Yes/No) No
  • Secrets/tokens handling changed? (Yes/No) No
  • New/changed network calls? (Yes/No) No
  • Command/tool execution surface changed? (Yes/No) No
  • Data access scope changed? (Yes/No) No
  • If any Yes, explain risk + mitigation:

Repro + Verification

Environment

  • OS: macOS
  • Runtime/container: local dev checkout
  • Model/provider: OpenAI / OpenAI Codex auth setup copy
  • Integration/channel (if any): onboarding/configure auth-choice UI text
  • Relevant config (redacted): default plugin/provider

Steps

  1. Inspect current OpenAI auth choice definitions in extensions/openai/openai-provider.ts and extensions/openai/openclaw.plugin.json.
  2. Verify the old hint still implied a combined OpenAI/Codex auth grouping.
  3. Update the hint text and run the repo checks triggered by commit hooks.

Expected

  • OpenAI setup text should clearly indicate API-key-only setup and not imply that Codex OAuth is included.

Actual

  • The old hint still suggested a combined OpenAI/Codex auth grouping until this copy change.

Evidence

Attach at least one:

  • Failing test/log before + passing after
  • Trace/log snippets
  • Screenshot/recording
  • Perf numbers (if relevant)

Human Verification (required)

What you personally verified (not just CI), and how:

  • Verified scenarios: Confirmed the stale hint existed on current main; updated both source locations so the displayed hint is now explicit about API key vs separate Codex OAuth.
  • Edge cases checked: Verified both the manifest and provider code path use the updated wording so they stay aligned.
  • What you did not verify: I did not run the full interactive onboarding flow end-to-end in the UI after the copy change.

Review Conversations

  • I replied to or resolved every bot review conversation I addressed in this PR.
  • I left unresolved only the conversations that still need reviewer or maintainer judgment.

If a bot review conversation is addressed by this PR, resolve that conversation yourself. Do not leave bot review conversation cleanup for maintainers.

Compatibility / Migration

  • Backward compatible? (Yes/No) Yes
  • Config/env changes? (Yes/No) No
  • Migration needed? (Yes/No) No
  • If yes, exact upgrade steps:

Risks and Mitigations

List only real risks for this PR. Add/remove entries as needed. If none, write None.

  • Risk: The new wording could still be considered incomplete if maintainers want stronger separation than a hint-only fix.
  • Mitigation: Kept the change intentionally narrow and copy-only, so it improves clarity without changing behavior or making broader UX assumptions.

Changed files

  • extensions/openai/openai-provider.ts (modified, +1/-1)
  • extensions/openai/openclaw.plugin.json (modified, +1/-1)
RAW_BUFFERClick to expand / collapse

Summary

OpenClaw currently makes it hard to understand how ChatGPT Plus / ChatGPT OAuth, Codex OAuth, and OpenAI API key relate when using GPT-5.4.

Two user-facing problems keep causing confusion:

  1. Users cannot tell whether ChatGPT Plus entitlement can be used directly for openai/gpt-5.4, or only indirectly via openai-codex/gpt-5.4.
  2. The usable quota/limits through openai-codex/gpt-5.4 appear significantly tighter than the ChatGPT web experience, but this difference is not clearly explained in OpenClaw UX/docs.

Current Behavior

  • openai/gpt-5.4 requires an OpenAI API key.
  • ChatGPT / Codex OAuth maps to openai-codex/gpt-5.4.
  • In practice, users often expect their ChatGPT Plus / ChatGPT subscription to behave like “GPT-5.4 access” in general, and do not understand why:
    • it does not grant direct API access for openai/gpt-5.4
    • the effective limits through openai-codex/gpt-5.4 may feel much smaller than the ChatGPT website/app

Why This Is Confusing

From a user perspective, “I already pay for GPT-5.4 in ChatGPT” naturally sounds like it should map to broad GPT-5.4 access in OpenClaw.

Instead, the actual model/auth matrix is more like:

  • ChatGPT / Codex OAuth -> openai-codex/gpt-5.4
  • OpenAI API key -> openai/gpt-5.4

That distinction exists, but it is not obvious enough during onboarding, model selection, or troubleshooting.

Requested Improvements

  1. Make the auth path explicit in docs and UI

    • Clearly explain that ChatGPT Plus / ChatGPT OAuth does not become direct OpenAI API permission.
    • Clearly explain that OAuth access currently maps to openai-codex/gpt-5.4, while openai/gpt-5.4 is API-key only.
  2. Document quota/limit differences clearly

    • If openai-codex/gpt-5.4 has different limits from the ChatGPT web app, say so directly.
    • If this is expected/provider-side behavior, document that expectation.
    • If this is not expected, please label it as a bug/regression.
  3. Improve model picker / onboarding wording

    • Example wording:
      • openai-codex/gpt-5.4 = ChatGPT/Codex OAuth route
      • openai/gpt-5.4 = OpenAI API key route
    • Ideally show a short note about likely billing/limit source for each route.
  4. Troubleshooting guidance

    • Add a FAQ entry for users asking:
      • “Why can ChatGPT web use GPT-5.4 but OpenClaw says I need an API key?”
      • “Why does Codex OAuth hit tighter limits than ChatGPT web?”

Related

  • Possibly related to #30055 (explicit OAuth vs API key routing)

Goal

This request is mainly about reducing user confusion and making quota/auth behavior legible, even if the underlying provider limitations cannot be changed.

extent analysis

TL;DR

Update OpenClaw documentation and UI to clearly explain the distinction between ChatGPT Plus/Codex OAuth and OpenAI API key authentication paths for GPT-5.4 access.

Guidance

  • Clarify in documentation that ChatGPT Plus/Codex OAuth does not grant direct access to openai/gpt-5.4 and instead maps to openai-codex/gpt-5.4.
  • Document quota and limit differences between openai-codex/gpt-5.4 and the ChatGPT web app, and label any unexpected behavior as a bug.
  • Improve model picker and onboarding wording to clearly indicate the authentication route and likely billing/limit source for each model.
  • Add troubleshooting guidance, including FAQ entries, to address common user questions about authentication and quota limits.

Example

A possible example of improved model picker wording:

* `openai-codex/gpt-5.4`: ChatGPT/Codex OAuth route (uses ChatGPT Plus/Codex OAuth credentials)
* `openai/gpt-5.4`: OpenAI API key route (requires separate OpenAI API key)

Notes

The goal of these changes is to reduce user confusion and make quota and authentication behavior more transparent, even if underlying provider limitations cannot be changed.

Recommendation

Apply workaround: Update documentation and UI to clearly explain authentication paths and quota limits, as this will help reduce user confusion and improve overall user experience.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING