Usage Tracking
Automate-E tracks token usage and calculates costs for every Claude API call.
How It Works
After each Claude API response, the usage tracker records:
- Input tokens -- tokens sent to Claude (system prompt + conversation history + tool results)
- Output tokens -- tokens generated by Claude (text response + tool calls)
- Model -- which model was used (determines pricing)
- Timestamp -- when the call was made
Per-Model Pricing
Cost calculation uses per-model token pricing:
| Model | Input (per 1M tokens) | Output (per 1M tokens) |
|---|---|---|
claude-haiku-4-5-20251001 |
$0.80 | $4.00 |
claude-sonnet-4-20250514 |
$3.00 | $15.00 |
claude-opus-4-20250514 |
$15.00 | $75.00 |
Note
Pricing is hardcoded in usage.js. Update when Anthropic changes rates.
Cost Calculation
For each API call:
cost = (input_tokens / 1_000_000 * input_price) +
(output_tokens / 1_000_000 * output_price)
Example: A Book-E message using Haiku with 2,000 input tokens and 500 output tokens:
cost = (2000 / 1_000_000 * 0.80) + (500 / 1_000_000 * 4.00)
= 0.0016 + 0.002
= $0.0036
Dashboard Integration
The usage tracker feeds data to the dashboard via WebSocket:
- Per-message cost shown next to each response
- Cumulative daily/monthly totals on the dashboard overview
- Token breakdown (input vs output) for optimization
Optimization Tips
| Strategy | Impact |
|---|---|
| Use Haiku for routine tasks | 18x cheaper than Opus |
Keep lore entries concise |
Reduces input tokens per message |
Set appropriate conversationRetention |
Shorter history = fewer input tokens |
Use messageExamples sparingly |
Each example adds to input tokens |