Low Key Studio · open source · runs locally
Where did my Codex / Claude Code quota go?
QuotaTrace reads the session logs already on your machine and lines up your plan's quota percentage with the tokens behind every call. You see which quota jumps were out of line with your work, and what actually burned your tokens.
npx github:lowkey-studio/quotatrace --since 30d
No terminal? Run it in your browser: pick your .codex/sessions or .claude/projects folder, and the analysis runs inside the page with nothing uploaded.
Node 18+, no dependencies, nothing uploaded. Writes quotatrace-report.md with counts only (hashed session ids, no prompts, paths or file names), so you can attach it to an issue or a support ticket. Source: github.com/lowkey-studio/quotatrace
What the report shows
- Quota points per million tokens, per window. Codex logs
used_percentwith each call. Comparable work should cost comparable points per million tokens; a sudden multiple on the same plan and model points to metering, not you. - Quota jumps out of line with the tokens behind them, with UTC timestamps. Jumps with zero local calls point to another device, a cloud task or background use.
- Where the tokens went: context replay, big-context calls that produced almost nothing, sleep/wait polling, and the tool outputs that were re-sent on every later call (size × times re-sent).
- Claude Code: cache writes vs reads, cache rebuilt after a pause, subagent share, and the tool results that were re-read the most. Claude Code keeps your plan's usage percentage out of its transcripts, so
quotatrace meter installadds a status line hook that records it (your existing status line keeps working). After that you get quota points per 1M tokens for Claude Code too.
Example rows from the sample report (synthetic logs):
| window | plan | quota used | tokens | % per 1M tokens |
|---|---|---|---|---|
| weekly | pro | 12% → 27% | 16.69M | 0.90 |
Quota jumps out of line
| time (UTC) | jump | tokens since last step | expected |
|---|---|---|---|
| 2026-10-10 15:20 | +9% | 123,300 | +0.1% |
Where the tokens went
- Waiting and polling: 40 calls (22.2%) followed a sleep / wait poll, costing 4.77M tokens (28.6%).
- One 60,000-token command output was re-sent 109 times: 6.54M input tokens.
Want a person to read it? QuotaTrace Review, $19
Email your quotatrace-report.md to [email protected] after paying. Within 6 hours you get back:
- the specific causes of your quota or token burn, read from your numbers;
- ranked fixes for your setup (AGENTS.md / CLAUDE.md, command output, polling, session and cache habits, subagents);
- if the report shows quota jumps out of line with your tokens, a ready-to-send support ticket that cites the timestamps and token counts.
See a sample review (written from the synthetic sample report).
The report holds counts only, so you never send code or prompts. If your report can't support any finding, you get a full refund.
Prefer crypto? Send 19 USDC on Base, or 19 USDC/USDT on Arbitrum One, to 0xB442daCAE60C07bb5dB8f33C8a6c1572d8F24e83 and email the transaction hash with your report.
Card checkout via Whop. Reviews are delivered by email from [email protected]. Orders for this round are accepted until 2026-10-16 02:50 UTC, so every order can be delivered within its 6 hours.
Contribute to the public quota dataset
npx github:lowkey-studio/quotatrace --share prints a link to a prefilled GitHub issue holding only per-window aggregates (date, window, plan, points used, thousands of tokens). You review it before posting. Pooled, these show what a quota point buys per plan and model, and when that changed. See submissions.
Caveats
Only calls logged on the machine you run it on are counted. Codex reports quota in whole points, so short windows are noisy. Tool-output sizes are estimated at 4 characters per token. Claude Code does not log your plan's usage percentage, so its section reports tokens. QuotaTrace is not affiliated with OpenAI or Anthropic.
Contact: [email protected] · public visit counter