The Analytics view provides comprehensive cost tracking, model comparison, efficiency metrics, and a sortable task performance table. Access it from the board toolbar (TrendingUp icon).
Cost Summary
Four stat cards at the top show:
| Metric | Description |
|---|
| Total Cost | Cumulative API cost across all tasks |
| Total Tokens | Input + output tokens with breakdown |
| Avg Cost/Task | Average cost per task with usage |
| Avg Tokens/Task | Average token consumption |
Model Comparison
A bar chart comparing models by cost and token usage:
- Each model (Haiku, Sonnet, Opus) shows task count, total tokens, and total cost
- Color-coded by model: green (Haiku), blue (Sonnet), purple (Opus)
- Helps identify which model is most cost-effective for your workload
Cost Trend
Daily cost chart showing spending over time. Hover for exact values per day (cost, tokens, task count).
Efficiency Metrics
Four efficiency cards:
| Metric | Description |
|---|
| Throughput | Tasks completed per hour of work time |
| Success Rate | Percentage of tasks reaching done/testing vs total |
| Cache Rate | Prompt cache hit percentage (higher = cheaper) |
| Avg Turns | Average conversation turns per task |
- Sort by: Cost (default), Tokens, or Duration — click column headers
- Columns: Task key, title, model, duration, tokens, cost
- Top 3 most expensive tasks are highlighted with a warning icon
- Paginated to 20 rows
Sort by cost descending to identify expensive tasks. Consider using a cheaper model (Haiku) for simple tasks like docs or chores.
