Fix Claude 0/0 usage with multi-source limits + auto-refresh#29
Merged
Conversation
The daemon's Claude path returned 0/0 because: - claude-plan-usage.json was past the 10-min TTL (no scheduler) - claude-statusline.json had no rate_limits (only set for some Pro auth states) - Token-counting fallback compared 1.2M+ tokens to a hardcoded 200k budget Changes: - Daemon now reads ~/.claude/plan-limits.json (statusline hook output) first, then the existing chrome-scrape cache, then statusline rate_limits, then the token-count fallback. Bumped chrome TTL to 60 min so partially stale data still surfaces real usage. - New scripts/claude-statusline-limits.py persists rate_limits from Claude Code statusLine into ~/.claude/plan-limits.json when present. - install.sh now schedules scripts/claude-plan-usage-cache.py via launchd every 10 min on macOS, keeping the chrome-scrape cache fresh when a claude.ai/usage tab is open. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 15e7060666
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Treat a hook payload missing either window as unknown and defer to the next source. Otherwise a partial payload would override a complete cached source with 100% remaining for the missing window. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
End-to-end staleness was dominated by the 10-minute scrape window; drop it so iOS sees fresh percentages within ~60s + 30s daemon poll + 30s app poll = ~2 minutes worst case. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
There was a problem hiding this comment.
Pull request overview
This PR improves Claude usage reporting in CodexMeter by adding a new “hook-written limits” data source, extending cache TTLs, and adding an auto-refresh mechanism so the daemon no longer falls back to misleading 0/0 usage when upstream sources are stale or missing fields.
Changes:
- Add
scripts/claude-statusline-limits.pyto persist Claude Coderate_limitsinto a normalized~/.claude/plan-limits.json. - Update the daemon to prefer hook limits first, and to use configurable TTLs for both hook and chrome-scrape caches.
- Extend the installer to add a macOS LaunchAgent that periodically refreshes the Claude plan usage cache.
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
scripts/claude-statusline-limits.py |
New helper script that normalizes and persists statusline rate limit data to a stable JSON file for the daemon to consume. |
scripts/claude-plan-usage-cache.py |
Minor cleanup (remove unused import) in the chrome-scrape cache script. |
install.sh |
Adds a macOS LaunchAgent to periodically run the chrome-scrape cache updater. |
daemon/codex-usage-daemon.py |
Adds hook-limits snapshot support, prioritizes it in local Claude usage resolution, and makes TTLs configurable. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| } | ||
|
|
||
|
|
||
| def normalized_limits(statusline: dict) -> dict | None: |
Comment on lines
+82
to
+92
| <key>StartInterval</key><integer>60</integer> | ||
| <key>RunAtLoad</key><true/> | ||
| <key>StandardOutPath</key><string>$HOME/Library/Logs/codexmeter-claude-scrape.log</string> | ||
| <key>StandardErrorPath</key><string>$HOME/Library/Logs/codexmeter-claude-scrape.err.log</string> | ||
| </dict> | ||
| </plist> | ||
| PLIST | ||
| launchctl unload "$SCRAPE_PLIST" >/dev/null 2>&1 || true | ||
| launchctl load "$SCRAPE_PLIST" | ||
| echo "Installed LaunchAgent: $SCRAPE_PLIST (every 60s)" | ||
| fi |
Comment on lines
982
to
+986
| def local_claude_activity_snapshot() -> UsageSnapshot: | ||
| hook_snapshot = claude_hook_limits_snapshot() | ||
| if hook_snapshot is not None: | ||
| return hook_snapshot | ||
|
|
4 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
s:0, w:0because all three data sources had become inert: the chrome-scrape cache was past its 10-min TTL with no scheduler refreshing it, the Claude Code statusline JSON had norate_limitsfield (only emitted under some Pro auth states), and the token-count fallback hit 100% used against a hardcoded 200k daily budget.~/.claude/plan-limits.json(hook output) →claude-plan-usage.json(chrome) →claude-statusline.json(rate_limits) → token-count fallback. Chrome TTL bumped to 60 min so slightly stale Pro data still surfaces over the 0% fallback.scripts/claude-statusline-limits.pypersistsrate_limitswhenever Claude Code statusline emits them.install.shnow installs a second LaunchAgent that runs the chrome-scrape every 10 min, so the cache stays fresh as long as a claude.ai/usage tab is open in Chrome.Test plan
pytest -q(11 passed)ruff check daemon/ scripts/cleanhttp://localhost:9596/usage→ returns reals/wpercentages ({"s":57,"sr":197,"w":96,"wr":2127,"st":"5h 43% used · 7d 4% used"})🤖 Generated with Claude Code