Documentation

Quickstart

AtlasFlux exposes an OpenAI-compatible API at https://api.atlasflux.my. This page is a frontend preview; full endpoint reference will follow with the backend.

Getting started

  1. 1

    Create an account

    Sign in with your AtlasFlux account. The same authentication is shared with AtlasFlux AI, but your API balance is separate.

  2. 2

    Generate an API key

    Create a key in the dashboard. Store it securely, it is shown only once. Use development keys while testing.

  3. 3

    Top up your balance

    Add prepaid credit in RM. All usage is deducted from this balance with no monthly commitment.

  4. 4

    Make your first request

    Call the responses endpoint with model atlasflux/nenas-flash. Streaming, reasoning and web search are enabled through request parameters.

Authentication

Authenticate with a bearer token. Keep keys on the server and never expose them in client code.

bash
curl https://api.atlasflux.my/v1/responses \
  -H "Authorization: Bearer $ATLASFLUX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "atlasflux/nenas-flash",
    "input": "Hello, explain the AtlasFlux API in one sentence"
  }'

Code examples

The API is compatible with standard OpenAI-style clients.

javascript
const res = await fetch("https://api.atlasflux.my/v1/responses", {
  method: "POST",
  headers: {
    "Authorization": `Bearer ${process.env.ATLASFLUX_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "atlasflux/nenas-flash",
    input: "Summarize this document",
    reasoning: { effort: "medium" },
  }),
});
const data = await res.json();
python
import requests

resp = requests.post(
    "https://api.atlasflux.my/v1/responses",
    headers={"Authorization": f"Bearer {ATLASFLUX_API_KEY}"},
    json={
        "model": "atlasflux/nenas-flash",
        "input": "Summarize this document",
        "reasoning": {"effort": "medium"},
    },
)
data = resp.json()

Key concepts

One model ID

Requests use atlasflux/nenas-flash. AtlasFlux routes to the best upstream model and you never manage providers.

Reasoning

Set reasoning.effort to low, medium or high. Reasoning tokens are billed at the output rate.

Web search

Enable search with modes off, auto or on, and depth from instant to reasoning. Content extraction is billed per page.

Billing

Prepaid balance in MYR. Spend is visible in real time and optional per-key monthly limits help control cost.

Ready to build?

Generate an API key and make your first request from the dashboard.

Get API key