Route and meter your agents' LLM calls at the edge.
steerz runs on Cloudflare next to your agents. Point Claude Code, Codex or any Anthropic or OpenAI SDK at it and every call is priced, logged and capped, using your own provider keys.
Connect your agents
Models
Ask for any of these by its provider id and steerz serves it as is. Any other model name goes to .
Smoke test
Runs from this browser through steerz to your providers: health, your key, free token counts, then one tiny real completion per provider (a few output tokens on the cheapest model).
Live
updating every 2 sEvery metered call your org makes, from every region it has used. Token counting is free and not counted.
Requests / sec
–
last 10 s
In flight
–
open now
Requests
–
last hour
Duration p50
–
last hour
Spend
–
all time, at list price
Requests per second, last minute
60 s agonow