-
Notifications
You must be signed in to change notification settings - Fork 819
Codex usage fetch fails on plan_type=prolite in v0.20 #709
Copy link
Copy link
Open
Labels
acceptedConfirmed backlog item or verified open bugConfirmed backlog item or verified open bugarea:usage-accuracyUsage math, reset windows, plan parsing, cost/token accuracyUsage math, reset windows, plan parsing, cost/token accuracybugSomething isn't workingSomething isn't workingpriority:highHigh priority: confirmed serious bug or blockerHigh priority: confirmed serious bug or blockerprovider:codexIssue specific to Codex/OpenAIIssue specific to Codex/OpenAI
Metadata
Metadata
Assignees
Labels
acceptedConfirmed backlog item or verified open bugConfirmed backlog item or verified open bugarea:usage-accuracyUsage math, reset windows, plan parsing, cost/token accuracyUsage math, reset windows, plan parsing, cost/token accuracybugSomething isn't workingSomething isn't workingpriority:highHigh priority: confirmed serious bug or blockerHigh priority: confirmed serious bug or blockerprovider:codexIssue specific to Codex/OpenAIIssue specific to Codex/OpenAI
Hi — CodexBar is failing to load Codex usage/rate limit data because the usage response now returns a
plan_typevalue ofprolite, which appears to be unsupported by the current decoder.Version
What happens
CodexBar shows:
Important detail
The server response itself indicates the request succeeded and the account is allowed:
{ "plan_type": "prolite", "rate_limit": { "allowed": true, "limit_reached": false } }So this looks like a client-side enum/schema mismatch, not an auth or rate-limit failure.
Full response body
{ "plan_type": "prolite", "rate_limit": { "allowed": true, "limit_reached": false, "primary_window": { "used_percent": 12, "limit_window_seconds": 18000, "reset_after_seconds": 8581, "reset_at": 1776111121 }, "secondary_window": { "used_percent": 2, "limit_window_seconds": 604800, "reset_after_seconds": 569914, "reset_at": 1776672455 } }, "code_review_rate_limit": null, "additional_rate_limits": [ { "limit_name": "GPT-5.3-Codex-Spark", "metered_feature": "codex_bengalfox", "rate_limit": { "allowed": true, "limit_reached": false, "primary_window": { "used_percent": 0, "limit_window_seconds": 18000, "reset_after_seconds": 18000, "reset_at": 1776120541 }, "secondary_window": { "used_percent": 0, "limit_window_seconds": 604800, "reset_after_seconds": 604800, "reset_at": 1776707341 } } } ], "credits": { "has_credits": false, "unlimited": false, "overage_limit_reached": false, "balance": "0", "approx_local_messages": [0, 0], "approx_cloud_messages": [0, 0] }, "spend_control": { "reached": false }, "promo": null }Expected
CodexBar should handle
plan_type=prolitegracefully and continue showing usage data.Actual
CodexBar fails decoding the response and reports a Codex connection error.
Guess
Likely just needs support for the new
proliteplan enum (or a safer unknown/fallback case).