Usage Tracker Skill

Track AI model usage and generate cost reports across all providers.

Overview

This skill tracks AI API calls and generates cost reports. It supports:

• Multiple providers: MiniMax, Anthropic, OpenAI, Google, xAI, Voyage, Mistral
• Cost calculation based on current pricing
• Daily, weekly, and monthly reports
• Filtering by model and task type

Tools

log_usage

Log an AI API call for tracking.

Parameters:

• provider (required): Provider name (minimax-portal, anthropic, openai, google, xai, voyage, mistral)
• model (required): Model identifier (e.g., MiniMax-M2.5, gpt-4o, claude-3-opus)
• input_tokens (required): Number of input tokens
• output_tokens (required): Number of output tokens
• task_type (optional): Type of task (coding, reasoning, general, writing, analysis, chat)

usage_report

Generate a usage and cost report.

Parameters:

• period (required): Time period (daily, weekly, monthly)
• model_filter (optional): Filter by model name
• task_filter (optional): Filter by task type

realtime_report

Get real-time usage report for the last 24 hours. No parameters required.

Usage Examples

Log a usage call:
log_usage(provider="minimax-portal", model="MiniMax-M2.5", input_tokens=1500, output_tokens=800, task_type="general")

Get daily report:
usage_report(period="daily")

Get weekly report:
usage_report(period="weekly", model_filter="gpt")

Get real-time report:
realtime_report()

Pricing (USD per 1M tokens)

Telegram Commands

• /usage - Get real-time usage report (last 24 hours)
• /cost - Get real-time usage report (last 24 hours)

Storage

Logs are stored in: usage-logs/api-cells.jsonl Reports can be generated on-demand using the tools above.