Every model costs its provider’s list price plus a 5.5% platform fee. The formula is on the pricing page, and every request shows its tokens and cost.
One API for every frontier model.
Point any OpenAI-compatible SDK at a single endpoint and reach authorized model providers. You pay each provider’s public list price plus a published 5.5% fee — nothing hidden in credits or exchange rates.
Pay as you go · No subscription · Prompts are not stored by default
Why AstrelayBuilt like infrastructure. Priced like a utility.
The boring parts done properly: clear prices, clean supply, private data, and a gateway that stops before it overspends.
Each upstream is a direct provider contract, an authorized cloud reseller program, or self-hosted open-weight models. Upstreams without an authorization record are refused in code.
We log metadata — time, model, token counts and status — not the content of your prompts or responses.
Give every service its own key with a spend limit, expiry, model allow-list and IP allow-list.
Chat completions, responses and streaming in the OpenAI format, plus Anthropic- and Gemini-format requests through the same gateway.
If a price changes or an upstream budget runs out, requests stop before they leave our servers. Billed requests are never retried automatically.
From key to first request in minutes.
Set a spend limit and an expiry in the console, then copy the key.
Point your existing OpenAI SDK at https://aidgu.com/v1. No other code changes.
Usage logs show model, tokens and cost per request, ready to export for your books.
curl https://aidgu.com/v1/chat/completions \
-H "Authorization: Bearer $ASTRELAY_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "<model-id>", "messages": [{"role": "user", "content": "Hello"}]}'import os
from openai import OpenAI
client = OpenAI(
base_url="https://aidgu.com/v1",
api_key=os.environ["ASTRELAY_API_KEY"],
)
reply = client.chat.completions.create(
model="<model-id>", # any model from the pricing page
messages=[{"role": "user", "content": "Hello"}],
)
print(reply.choices[0].message.content)import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://aidgu.com/v1",
apiKey: process.env.ASTRELAY_API_KEY,
});
const reply = await client.chat.completions.create({
model: "<model-id>",
messages: [{ role: "user", content: "Hello" }],
});
console.log(reply.choices[0].message.content);List price plus 5.5%. That’s the whole formula.
No credits that expire, no markup hidden in exchange rates. Volume terms and monthly invoicing are available for enterprise.
Your price = provider list price × (1 + 5.5%)
Good to know
Which models can I use?
Every model on the pricing page, each served through an upstream we’re authorized to use. A model goes live only after its authorization record is on file, so the list grows as agreements are signed.
Where is Astrelay available?
Available only where each upstream provider supports service. Not available in mainland China, Hong Kong or Macau.
Do you store my prompts?
No. We keep the request metadata needed for billing and support. Prompt and response content isn’t written to our logs by default.
How is billing calculated?
Each request is metered in tokens and charged at that model’s list price plus the 5.5% platform fee. Usage logs show the tokens and cost of every request.
Can I bring my own provider key?
Bring-your-own-key routing is on our roadmap. Email [email protected] if you need it now.
Is there an enterprise plan?
Yes — monthly invoicing, a custom data processing agreement and volume terms. Email [email protected].
Ship on every model with one integration.
Request access and we’ll set up your workspace, usually within one business day.