Skip to content

feat: track GPT model usage and reference costs - #1

Closed
679215 wants to merge 1 commit into
Littps:mainfrom
679215:codex/gpt-model-cost-records
Closed

679215 wants to merge 1 commit into
Littps:mainfrom
679215:codex/gpt-model-cost-records

Conversation

@679215

@679215 679215 commented Nov 3, 2026

Copy link
Copy Markdown

Summary

  • Add per-model GPT token totals, request counts, cached/input/output breakdowns, and reference costs across request, turn, session, and the existing time ranges.
  • Add validated public price lookup with caching; unsupported models, missing prices, and ambiguous mappings stay unknown rather than being guessed or double-counted.
  • Add model list and single-model details in the UI, model-usage history, and schema v5 while retaining auth v2.
  • No account/subscription quota experiments, no message-body or credential reads, and no npm runtime dependency.

Validation

  • npm run check — passed (syntax, ESLint, Prettier).
  • npm test — passed (119 tests; snapshot check passed).
  • npm run test:auth-ui — passed (9 tests).
  • Live checks are not green in this environment: the selected Codex pages did not have the panel injected (verify-live 26/29; test:live 17 failed / 2 skipped). These require a deployed monitor and live Codex page; no deployment is included in this PR.
  • Real-machine golden corpus and Node.js 22 CI remain pending.

Draft PR for upstream review; not merged.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants