Infer
Open console
Inference routing for your API keys

Your models.
One route in.

Bring your provider accounts together and send requests through one API key. Infer chooses an available account for the exact model you ask for, with limits you control.

Get started with GitHubUse your own upstream credentials
Request examplePOST /v1/chat/completions
{
  "model": "deepseek/deepseek-flash",
  "messages": [
    { "role": "user", "content": "Hello" }
  ]
}
ChatResponsesMessages

One workspace

Everything around the request, in one place.

Routes that use your accounts

Connect multiple OpenAI, Anthropic, Gemini, DeepSeek, or Kilo credentials. Pick the exact models each account can serve.

Limits enforced in Redis

A Lua decision selects an eligible account on every request and applies your per-minute and monthly caps atomically.

Trace every routed call

Open a request to inspect headers, payloads, provider responses, timing, and errors. Credential values are redacted.