/aqe-costs

name: aqe-costs
description: Display inference cost analysis and savings from local vs cloud providers

AQE Inference Costs

Display comprehensive inference cost analysis showing local vs cloud inference costs and estimated savings.

Usage

aqe costs [options]
# or
/aqe-costs [options]

Options

| Option | Type | Default | Description | |--------|------|---------|-------------| | `--period` | string | `24h` | Time period: 1h, 24h, 7d, 30d, all | | `--provider` | string | - | Filter by provider: ruvllm, anthropic, openrouter, openai | | `--format` | string | `text` | Output format: text, json | | `--detailed` | boolean | `false` | Show detailed per-request breakdown | | `--reset` | boolean | `false` | Reset cost tracking data |

Examples

Basic Cost Report

aqe costs

Displays cost summary for the last 24 hours with savings analysis.

Weekly Cost Analysis

aqe costs --period 7d

Shows cost trends and savings over the past 7 days.

Provider-Specific Costs

aqe costs --provider ruvllm

Displays costs for local ruvllm inference only.

Detailed Breakdown

aqe costs --detailed

Shows per-request cost breakdown with agent and task attribution.

JSON Export for Dashboards

aqe costs --format json > costs.json

Exports cost data in JSON format for integration with monitoring dashboards.

Reset Cost Data

aqe costs --reset

Clears all tracked cost data (useful for testing or new billing periods).

Integration with Claude Code

Cost Monitoring Agent

// Use Claude Code's Task tool for cost monitoring
Task("Monitor inference costs", `
  Analyze AQE inference costs and provide recommendations:
  - Check cost trends over the past 24 hours
  - Identify high-cost agents or tasks
  - Calculate savings from local inference
  - Recommend optimizations to reduce cloud costs

  Store findings in memory: aqe/costs/analysis/{timestamp}
`, "qe-quality-gate")

Automated Cost Reporting Workflow

// Daily cost report generation
[Single Message]:
  Task("Generate cost report", "Create daily inference cost summary", "qe-quality-gate")
  Task("Analyze cost trends", "Identify cost optimization opportunities", "qe-quality-gate")

  TodoWrite({ todos: [
    {content: "Fetch cost data from tracker", status: "in_progress", activeForm: "Fetching data"},
    {content: "Calculate savings metrics", status: "in_progress", activeForm: "Calculating savings"},
    {content: "Generate recommendations", status: "pending", activeForm: "Generating recommendations"},
    {content: "Store report in memory", status: "pending", activeForm: "Storing report"}
  ]})

Expected Outputs

Text Format (Default)

Inference Cost Report
====================

Period: 2025-12-15T00:00:00Z to 2025-12-15T23:59:59Z

Overall Metrics:
  Total Requests: 1,248
  Total Tokens: 3,456,789
  Total Cost: $5.2340
  Requests/Hour: 52.0
  Cost/Hour: $0.2181

Cost Savings Analysis:
  Actual Cost: $5.2340
  Cloud Baseline Cost: $18.7650
  Total Savings: $13.5310 (72.1%)
  Local Requests: 892 (71.5%)
  Cloud Requests: 356 (28.5%)

By Provider:
  🏠 ruvllm:
    Requests: 892
    Tokens: 2,234,567
    Cost: $0.0000
    Avg Cost/Request: $0.000000
    Top Model: meta-llama/llama-3.1-8b-instruct

  ☁️ anthropic:
    Requests: 245
    Tokens: 891,234
    Cost: $4.5678
    Avg Cost/Request: $0.018644
    Top Model: claude-sonnet-4-6

  ☁️ openrouter:
    Requests: 111
    Tokens: 330,988
    Cost: $0.6662
    Avg Cost/Request: $0.006002
    Top Model: meta-llama/llama-3.1-70b-instruct

JSON Format

{
  "timestamp": "2025-12-15T23:59:59Z",
  "period": {
    "start": "2025-12-15T00:00:00Z",
    "end": "2025-12-15T23:59:59Z"
  },
  "overall": {
    "totalRequests": 1248,
    "totalTokens": 3456789,
    "totalCost": 5.234,
    "requestsPerHour": 52.0,
    "costPerHour": 0.2181
  },
  "savings": {
    "actualCost": 5.234,
    "cloudBaselineCost": 18.765,
    "totalSavings": 13.531,
    "savingsPercentage": 72.1,
    "localRequestPercentage": 71.5,
    "cloudRequestPercentage": 28.5,
    "localRequests": 892,
    "cloudRequests": 356,
    "totalRequests": 1248
  },
  "byProvider": {
    "ruvllm": {
      "provider": "ruvllm",
      "providerType": "local",
      "requestCount": 892,
      "inputTokens": 1489711,
      "outputTokens": 744856,
      "totalTokens": 2234567,
      "totalCost": 0,
      "avgCostPerRequest": 0,
      "topModel": "meta-llama/llama-3.1-8b-instruct",
      "modelCounts": {
        "meta-llama/llama-3.1-8b-instruct": 892
      }
    },
    "anthropic": {
      "provider": "anthropic",
      "providerType": "cloud",
      "requestCount": 245,
      "inputTokens": 594156,
      "outputTokens": 297078,
      "totalTokens": 891234,
      "totalCost": 4.5678,
      "avgCostPerRequest": 0.018644,
      "topModel": "claude-sonnet-4-6",
      "modelCounts": {
        "claude-sonnet-4-6": 187,
        "claude-haiku-4-5-20251001": 58
      }
    },
    "openrouter": {
      "provider": "openrouter",
      "providerType": "cloud",
      "requestCount": 111,
      "inputTokens": 220659,
      "outputTokens": 110329,
      "totalTokens": 330988,
      "totalCost": 0.6662,
      "avgCostPerRequest": 0.006002,
      "topModel": "meta-llama/llama-3.1-70b-instruct",
      "modelCounts": {
        "meta-llama/llama-3.1-70b-instruct": 111
      }
    }
  }
}

Detailed Format

Inference Cost Report (Detailed)
================================

Period: 2025-12-15T00:00:00Z to 2025-12-15T23:59:59Z

Recent Requests (Last 20):

[2025-12-15T23:58:45Z] ruvllm/meta-llama/llama-3.1-8b-instruct
  Agent: qe-test-generator
  Tokens: 1,234