Changelog
What's new in Argus, in order.
v1.1.0Latest
July 22, 2026
Argus sources rates
Argus now computes every cost on its own backend from the signals your app already sends — and lets you override them when your deal is different.
- Server-side pricing. Stream the endpoint you called (or the model) plus units; Argus resolves the provider, the operation, and the price. Your app never needs our SKU names.
- Direct-provider LLM rates. Native calls to OpenAI, Anthropic, Gemini, DeepSeek, and Perplexity now price by token — not just OpenRouter-routed traffic.
- Custom rate cards. Override any unit rate, or set a bespoke plan — per request, or per workspace in the dashboard — so negotiated deals price correctly.
- Subscription overage. Once a plan's included allowance is spent, the excess is priced and tracked as spend on top of the monthly fee.
- ~90 per-operation rates. Across all 28 supported providers — search, enrichment, scraping, email, and inference.
v1.0.0
July 21, 2026
One eye for every API
Argus launches — spend, balances, and rate limits across every metered API you pay for, in one place.
- Three ways to connect. Paste a key to poll, wrap fetch to push per-call telemetry, or point an agent at the MCP server.
- MCP server. A read-only meter server, so the agents doing the spending can watch their own runway.
- Runway & alerts. Days-to-zero, burn rate, and rate-limit headroom — with an email before a meter runs dry.
- Provider catalog. Plans, unit prices, rate limits, and exhaustion behavior for every supported provider.
Building on Argus? Read the docs or start free.