Home / Cost analytics
Cost analytics
Live LLM spend, model breakdown, and cache effectiveness across the AIDP agent fleet. Data is pulled from the master agent's /cost/summary endpoint, which aggregates the agent_calls audit table in D1. Switch windows below; filter by user to scope spend to your own session history.
Total spend
—
window total
Calls
—
routed agent calls
Cache hit rate
—
hits / total
Your spend
—
filter by user above
Spend by model
Model breakdown
| Model | Calls | Tokens | Cost |
|---|
FinOps · forecast & infrastructure — from the Cost agent
LLM-spend forecast plus a modelled Cloudflare infrastructure estimate at public list prices, and FinOps signals. Estimates — tune in the agent.
Projected LLM / month
—
run-rate
Modelled infra / month
—
Workers · D1 · R2 · Vectorize · KV · AI
Projected total / month
—
LLM + infra (modelled)
Projected / year
—
LLM run-rate × 365
FinOps signals
Loading…
Modelled infra line items
| Item | Basis | Monthly |
|---|---|---|
| Loading… | ||
Cost by program — modelled: trace hops × avg cost/call
| Program | Hops | Est. cost |
|---|---|---|
| Loading… | ||
Recent calls
Filter by user above to see their recent call history.