model-usage

You can summarize per-model CodexBar usage by running a script that reads local cost JSON payloads and aggregates expenses by model. It identifies the current model and outputs a concise cost breakdown for monitoring spending.

Yes, you can filter CodexBar cost logs by applying optional day-range selection when running the script. This narrows the per-model cost aggregation to specific daily entries instead of summarizing all historical data.

The current model view shows the latest daily entry for the active model, while the all models view aggregates costs across every model in the logs. Both help identify which models are accumulating the highest usage expenses.

Yes, the usage summary supports output in both plain text and JSON formats. This allows you to easily integrate the aggregated per-model cost data into automated reporting pipelines or other downstream analytics workflows.

Yes, the script supports input from both file and stdin. This allows you to pipe CodexBar cost JSON payloads directly into the aggregator for per-model usage summarization without needing to save intermediate files.

Raw CodexBar local cost logs are difficult to compare because they are not aggregated by model. Summarizing these JSON payloads into a per-model usage breakdown makes it easier to identify expensive models and monitor costs.