One endpoint, every model.
Point the SDK you already use at one URL. Each call goes to the model you name, or the cheapest one that passes a check you set (a test, a schema, a judge), and every call is recorded with what it cost. Over a plain key you get routing, checks, one bill and the record.
# Keys come from the waiting list; this URL answers once you have one.
from openai import OpenAI
client = OpenAI(base_url="https://api.massrelay.ai/v1", api_key="YOUR_RELAY_KEY")
client.chat.completions.create(model="auto", messages=[{"role": "user", "content": "Hello"}])List price
Each model at its provider's list price. Fees are shown before they apply.
Your own key or model
Bring your own key or your own server; we route, check and record.
Open models we host
Open-weight models on our own servers, at our own price. Coming.
Build for a field
Start from AI by field, then build your own field assistant on the router.
Run a model on your own machine? Offer it to others, free for now
Talk to us.
Your company, its size and the first job you would hand to an assistant. We answer by email, nothing else.
Tell us your company, its size and the first job you would hand over. A person answers.