Cost control
Spend by agent, model, team and customer. Budgets that alert at 80% and stop at 100%. Forecasts on day five of what day thirty-one will look like.
LLM spend, last 30 days
$594.17
What you get
Token-accurate cost for every span, with cache hits and model pricing applied.
Soft limits alert, hard limits pause the agent. Per agent, team or workspace.
Hourly spend above 3× baseline pages the owner before it becomes a line item.
Month-end projections from current run rate, with the budget line drawn in.
See which models carry your spend and what moving a step to a smaller model saves.
Tag runs by team or customer and export monthly cost reports as CSV.
How it works
Every LLM span is priced from the model and token counts. No setup.
budget:
scope: agent/research-assistant
monthly_usd: 260
alert_at: 80%
hard_limit: 100% # pauses the agentPick a scope, a monthly amount, and what happens at 80% and 100%.
Fernline in #ai-ops
Research Assistant is spending 5.9× its hourly baseline. At this rate October lands at $1,054 against a $260 budget.
Alerts and forecasts land in Slack while there's still time to act.
Customers
“Budgets with a hard stop were the only way finance would let us put agents on the expensive models. Spend is down 30% anyway.”
FAQ
Can't find what you're looking for? Write to us and a human answers within a day.
Free for your first 100k spans. Set up in ten minutes, see your first trace in two.
No card needed · SOC 2 Type II · Cancel anytime