VSCode extension that automatically refreshes Codex state and pings the Codex API when rate limits reset.
Why
This extension automatically restarts rate limits by pinging Codex at the exact moment when cooldowns expire. This allows you to always fully utilize Codex limits, ensuring maximum usage efficiency without manual intervention.
How it works
The extension:
Reads Codex authentication from ~/.codex/auth.json
Uses the model from ~/.codex/config.toml or an optional VS Code setting override
Falls back to current ChatGPT-account-compatible Codex models if the preferred model stops working
Reads limit data from x-codex-* response headers
Supports both reset-after-seconds and reset-at header formats
Keeps a background refresh loop so the extension can recover if a scheduled timer was missed
Automatically sends a ping to Codex API at the moment of reset
This ensures you always have up-to-date rate limit information immediately after a reset occurs.
Prerequisites
Codex authentication (codex login)
Authentication stored in ~/.codex/auth.json
Preferred model usually stored in ~/.codex/config.toml
Configuration
codexPing.refreshIntervalSeconds - background refresh interval in seconds (default: 60)
codexPing.requestModel - optional model override for ping requests; leave empty to follow ~/.codex/config.toml
