llmtrace 30D SPEND$379.87 ANOM2 ACTIONS0 AGENT ACTIVEscan 1 NEXT·
DAILY SPEND · 30Dwhich deploy spiked the bill?
TRACKED KEYS3 keys · anthropic
KEY30D TREND$/DAYΔMODEL MIXσSTATUS
prod-frontend $8.45 ▲ 3.5× 62% haiku 5.2 HIGH
background-jobs $2.29 ━ flat sonnet · OK
internal-tools $1.92 ━ flat sonnet · OK
Autonomous Agent · Active scans 1 · last 02:17:10 UTC · next ·
Projected · prod-frontend $603/mo baseline run-rate $136/mo · this regression adds +$467/mo if left unfixed
Spend Anomalies
ANOMALY prod-frontend 2026-06-08
actual$20.85
baseline 7d$5.91
delta+$14.94
sigma5.2σ
calls830
ANOMALY prod-frontend 2026-06-07
actual$12.92
baseline 7d$4.68
delta+$8.24
sigma28.0σ
calls838
Ask the Agent
>
Investigate Any DashboardGemini Vision
Drop a screenshot of any LLM billing dashboard: Anthropic, OpenAI, a Grafana panel. Gemini reads the spend off the image, then the agent finds which deploy caused the spike.
● READY scan #1 last 02:17:10 UTC next · proxy :8080 MIT · llmtrace v0.2