Waiting for a sample.
A clearer view of
your AI usage.
Track Codex limits, credits, and API costs by hour.
Sign in with FilesUse your workspace passkey.
Your usage, by the hour.
Codex limits and recorded API activity, in one place.
Request log
Recent Codex requests, with model calls and tool steps. Times use UTC.
How times and credit estimates work
Credits are estimates from account balance changes, shared across model calls by token cost. Other account activity can affect them.
USD uses $0.052 per credit. Missing samples remain unknown. Tool rows exclude any separate provider charges.
Model times cover recorded output activity. They can exclude time before the first output. Tool times cover the call and return.
Step explanations use visible replies, tool names, and recorded status. They do not use internal reasoning.
Tool details show recorded events, including actions inside exec. Start-time grouping is an estimate when call IDs differ.
Tool event times can overlap wrapper times. Do not add them together. Missing events remain unavailable.
Tool content loads when expanded. Detected credentials are redacted. Sensitive sources and internal reasoning are excluded. Recording adds no model calls.
Latest reported balance
Estimated USD · recorded requests
Estimated USD · recorded requests
Hourly activity
Your rate: $260 per 5,000 credits ($0.052 per credit).
Observed
usage only. Gaps remain unknown.
API equivalent uses the recorded speed when available. Unknown speeds use Standard API rates. Credits cover account samples. Tokens cover this node, so totals can cover different activity.
How the API estimate works
Each response uses its recorded model, input, cached input, and output tokens. Long requests use higher rates.
Reasoning tokens count within output. Missing model prices or token details remain unpriced.
Fast mode uses twice the Standard API rate when recorded. Cache-write premiums, tool fees, media, regional fees, and tax are excluded.
This estimates an alternative cost. It is separate from recorded API spending below. View API rates.
View hourly numbers
| Hour · local time | Codex input | Cached input | Output | Credits used* | Credit cost (USD)* | Codex API equivalent (USD)* | API tokens | API cost* |
|---|
API breakdown
Selected time range · USD estimates
| Source | Model | Requests | Input tokens | Output tokens | Estimated cost |
|---|
Costs exclude unpriced requests, tools, media, taxes, and calls outside these sources. This is not a billing statement.
Data coverage
Codex tokens cover sessions on this node. Limits and credits reflect account samples.
API data covers OpenAI calls recorded by AP, Arena, and Genie. Copied AP staging rows count once.
API account credit balance and other providers are not connected. Older Genie records lack model and cache fields.
Rates checked 21 September 2026. Pricing · Codex limits
Dashboard alerts
Show an alert here when a threshold is reached.