LLM & Costs
Track every dollar spent on LLM API calls.
The LLM & Costs tab gives you complete visibility into your AI spending.
Cost cards
Total Cost
Total spend for the selected period
Total Tokens
Input + output tokens consumed
Cost Per Request
Average cost per API call
Top Spender
The agent or model costing the most
Cost by Model Over Time
An interactive line chart showing cost trends broken down by model:
Each model gets its own coloured line
Hover for exact daily costs
Toggle models on/off to compare
Time range selector (7D, 30D, 90D)
Cost breakdown
Table showing each model's:
Total cost
Token count (input/output split)
Request count
Avg cost per request
Table showing each agent's:
Total cost
Primary model used
Request count
Cost trend (increasing/decreasing)
Daily/weekly/monthly cost aggregation:
Bar chart of spend over time
Moving average trend line
Projected monthly cost
Cost Projections
Based on current usage patterns, the platform projects:
Monthly estimate — projected cost for the current month
Trend direction — increasing, stable, or decreasing
Anomaly alerts — unusual cost spikes flagged
Cost projections are estimates based on recent usage. Actual costs depend on usage patterns, which can change.
Tips for cost optimisation
Use cheaper models for simple tasks — route straightforward queries to Haiku or GPT-3.5
Monitor the Top Spender card — investigate if one agent dominates costs
Check token efficiency — agents with low efficiency may need prompt optimisation
Set cost alerts — configure budget thresholds in Settings
Last updated