api-token-usage-tracker

name	API Token Usage Tracker
description	"追踪和监控 LLM API Token 消耗，实时查看使用量、成本、配额。支持 OpenAI、Kimi、GLM 等多种 API。"

When to Use

Use this skill when you want to:

• Check how many tokens you've used today/this week/this month
• Monitor API costs and spending
• View remaining quotas
• Track token usage across different models
• Analyze usage patterns

Triggers:

• User says "token usage", "API usage", "token 消耗", "API 使用量"
• User asks "how many tokens", "用了多少 token"
• User mentions "API cost", "API 成本", "配额"
• User wants to "check usage", "查看使用量"

Quick Start

Check Today's Usage

查看今天的 token 使用量

Check Weekly Usage

查看本周的 API 使用情况

Check Costs

计算本月的 API 成本

Manual Run

python {workspace}/token-tracker/track-usage.py

Features

1. Real-time Tracking

Tracks token usage in real-time:

• Input tokens
• Output tokens
• Total tokens
• Cached tokens (if applicable)

2. Cost Calculation

Calculates costs based on:

• Model-specific pricing
• Input/output token rates
• Cache rates (if applicable)
• Currency conversion

3. Usage Analytics

Provides insights on:

• Usage trends (daily/weekly/monthly)
• Most used models
• Peak usage times
• Cost distribution

4. Quota Monitoring

Monitors quotas:

• Remaining tokens
• Rate limits
• Quota warnings

5. Multi-Provider Support

Supports multiple providers:

• OpenAI (GPT-4, GPT-3.5, etc.)
• Kimi (Moonshot)
• GLM (Zhipu AI)
• Custom providers

Configuration

Provider Config (config/providers.json)

{
  "providers": {
    "openai": {
      "name": "OpenAI",
      "api_key": "sk-...",
      "base_url": "https://api.openai.com/v1",
      "models": {
        "gpt-4": {
          "input_cost": 0.03,
          "output_cost": 0.06,
          "currency": "USD"
        },
        "gpt-3.5-turbo": {
          "input_cost": 0.0015,
          "output_cost": 0.002,
          "currency": "USD"
        }
      }
    },
    "kimi": {
      "name": "Kimi",
      "api_key": "sk-...",
      "base_url": "https://api.moonshot.cn/v1",
      "models": {
        "moonshot-v1-8k": {
          "input_cost": 0.012,
          "output_cost": 0.012,
          "currency": "CNY"
        }
      }
    }
  }
}

Usage Log Format (data/usage.json)

{
  "usage": [
    {
      "timestamp": "2024-03-25T09:00:00Z",
      "provider": "kimi",
      "model": "moonshot-v1-8k",
      "input_tokens": 1500,
      "output_tokens": 500,
      "total_tokens": 2000,
      "cached_tokens": 0,
      "cost_usd": 0.0024,
      "session_id": "abc123"
    }
  ]
}

Usage Report Example

# 📊 API Token 使用报告 - 2024-03-25

## 今日概览

| 指标 | 数值 |
|------|------|
| 总请求数 | 42 |
| 输入 Tokens | 125,000 |
| 输出 Tokens | 35,000 |
| 总 Tokens | 160,000 |
| 成本 (USD) | $3.45 |

## 分模型统计

| 模型 | 请求数 | 输入 | 输出 | 成本 |
|------|--------|------|------|------|
| GPT-4 | 15 | 50K | 10K | $2.10 |
| GPT-3.5 | 20 | 60K | 20K | $0.13 |
| Kimi | 7 | 15K | 5K | $0.22 |

## 分时段统计

09:00-12:00: ████████ 40,000 tokens 12:00-15:00: ██████ 30,000 tokens 15:00-18:00: ████████ 35,000 tokens 18:00-21:00: ████████ 35,000 tokens 21:00-24:00: ████ 20,000 tokens

## 成本趋势（最近 7 天）

03/19: $2.50 03/20: $3.20 03/21: $2.80 03/22: $4.10 03/23: $3.50 03/24: $3.80 03/25: $3.45 ← Today

## 配额警告

⚠️ **Kimi API**: 本月已使用 78% 配额
- 已用: 780,000 tokens
- 总配额: 1,000,000 tokens
- 剩余: 220,000 tokens

## 建议

1. 考虑使用 GPT-3.5 替代 GPT-4 以降低成本（简单任务）
2. Kimi API 接近配额上限，建议控制使用
3. 高峰时段 09:00-12:00，可考虑错峰使用

Commands

Check Usage

# 今日使用
python track-usage.py --today

# 本周使用
python track-usage.py --week

# 本月使用
python track-usage.py --month

# 自定义时间范围
python track-usage.py --start 2024-03-01 --end 2024-03-25

Check Costs

# 计算成本
python track-usage.py --calculate-cost

# 按模型分组
python track-usage.py --group-by model

# 按日期分组
python track-usage.py --group-by date

Export Reports

# 导出为 JSON
python track-usage.py --export json

# 导出为 CSV
python track-usage.py --export csv

# 导出为 Markdown
python track-usage.py --export markdown

Integration with OpenClaw

Automatic Logging

Automatically logs usage when OpenClaw makes API calls:

# In your OpenClaw config
{
  "hooks": {
    "usage-logger": {
      "enabled": true,
      "log_path": "{workspace}/token-tracker/data/usage.json"
    }
  }
}

Real-time Dashboard

View usage in real-time:

python dashboard.py --port 8080

Then open http://localhost:8080 in your browser.

Pricing Reference

OpenAI (as of 2024)

Model	Input	Output	Cache Read	Cache Write
GPT-4 Turbo	$0.01/1K	$0.03/1K	$0.003/1K	$0.005/1K
GPT-4	$0.03/1K	$0.06/1K	-	-
GPT-3.5 Turbo	$0.0005/1K	$0.0015/1K	-	-

Kimi (Moonshot)

Model	Input	Output
moonshot-v1-8k	¥0.012/1K	¥0.012/1K
moonshot-v1-32k	¥0.024/1K	¥0.024/1K
moonshot-v1-128k	¥0.06/1K	¥0.06/1K

GLM (Zhipu AI)

Model	Input	Output
glm-4	¥0.1/1K	¥0.1/1K
glm-4-flash	Free	Free

Advanced Features

Budget Alerts

Set budget limits and receive alerts:

{
  "budgets": {
    "daily": 5.00,
    "weekly": 30.00,
    "monthly": 100.00
  },
  "alerts": {
    "email": "[e-mail]",
    "webhook": "https://..."
  }
}

Cost Optimization Suggestions

Get AI-powered suggestions to reduce costs:

python track-usage.py --optimize

Comparative Analysis

Compare usage across periods:

python track-usage.py --compare last-week,this-week

Troubleshooting

No Usage Data Found

Error: No usage data available

Solution: Ensure usage logging is enabled in OpenClaw config.

API Key Invalid

Error: Cannot fetch quota information

Solution: Check API key in config/providers.json.

Incorrect Cost Calculation

Warning: Cost calculation may be inaccurate

Solution: Update pricing data in config file.

Best Practices

1. Check Daily: Monitor usage at the end of each day
2. Set Budgets: Configure budget alerts to avoid overspending
3. Optimize Models: Use cheaper models for simple tasks
4. Cache Tokens: Take advantage of caching when available
5. Review Weekly: Analyze weekly trends to optimize usage

Related Skills

• cost-optimizer - AI-powered cost optimization
• quota-manager - Manage and allocate API quotas
• usage-analytics - Advanced usage analytics

Future Enhancements

• Web dashboard with real-time charts
• Slack/Discord notifications
• Cost prediction based on usage patterns
• Multi-user support
• Budget sharing across teams