Plutus · LLM spend monitoring

Know your burn before the credits run out.

Plutus is a CLI dashboard that tracks API credit balances across every LLM provider in real time — burn-rate analysis, depletion forecasting, and usage alerts — so a stalled agent run is never a billing surprise.

<200ms
Balance query
6+
Providers tracked
7-day
Depletion forecast
<60s
Setup
What it does

Every provider's balance, in one view.

Live balance tracking

Real-time credit balances across all configured providers, refreshed hourly — one dashboard, no portal-hopping.

Burn-rate analysis

Spend velocity per provider and per project, so you see which workloads are eating the budget.

Depletion forecast

7-day predictive modeling tells you when you'll run dry — before an agent run stalls mid-task.

Usage alerts

Configurable thresholds fire before you hit a wall. No more silent 429s in production.

Custom providers

Add any provider in ~50 lines of Python. The plugin surface is small by design.

Perseus integration

Surface live balances inside agent context with the @plutus directive — financial awareness, in-band.

Anthropic LIVE OpenAI LIVE Google LIVE AWS Bedrock LIVE Azure OpenAI BETA Custom PLUGIN
One command

Set up in under a minute.

Point Plutus at your provider keys once. From then on, plutus status is your single source of truth for AI spend — locally, or surfaced into agent context through Perseus.

CLI dashboardHourly refreshSpend ledger
plutus status
$ plutus status # balances · refreshed 4m ago Anthropic $842.10 ~12d left OpenAI $310.55 ~6d left Google $48.20 ~2d left ⚠ burn: $71/day · alert at $50