model-usage

Summarize per-model usage and cost data from CodexBar cost logs.

23|10|Updated Feb 9, 2026
One-click install
npx skills add https://github.com/deepgram/dglabs-deepclaw --skill model-usage-deepgram
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: model-usage
Source: https://github.com/deepgram/dglabs-deepclaw/tree/main/skills/model-usage
Command: npx skills add https://github.com/deepgram/dglabs-deepclaw --skill model-usage-deepgram

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires python, and includes scripts (resource) components.

What problem does it solve?

This Skill provides a summary of per-model usage and cost data for Codex or Claude, helping users understand their model-level consumption and costs.

Core Features & Use Cases

  • Usage Cost Summary: Summarize per-model usage cost from CodexBar's local cost logs.
  • Current Model or All Models: Supports both current model usage/cost data and full model breakdowns.
  • Customization: Can provide a scriptable per-model summary from CodexBar cost JSON.

Quick Start

Run the script with the desired provider and mode:

python {baseDir}/scripts/model_usage.py --provider codex --mode current

Frequently Asked Questions about model-usage

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I analyze per-model usage and cost data from CodexBar logs?

To analyze per-model usage and cost data from CodexBar logs, run the provided Python script to parse local cost JSON files and summarize consumption metrics for Codex or Claude models. The script extracts usage metrics directly from the logs.

What is the difference between current model usage and full model breakdown in CodexBar?

Current model usage shows cost data for the active model, while full model breakdowns provide a comprehensive summary across all models. The script supports both modes to help you understand total model-level consumption.

Do I need Python and CodexBar CLI to retrieve model cost summaries?

Yes, you need Python installed to execute the analysis scripts, and the CodexBar CLI is required for data retrieval. These dependencies are necessary to access and parse the local cost logs for summary generation.

Can I customize the per-model summary output from CodexBar cost JSON?

Yes, the script provides a scriptable per-model summary extracted from CodexBar cost JSON. You can customize the execution through command-line arguments to specify the provider and mode for tailored cost analysis.

What's the best way to track Codex and Claude model consumption costs?

The best way to track Codex and Claude model consumption costs is by summarizing CodexBar's local cost logs. This approach provides per-model usage data directly from your existing cost tracking infrastructure.