Documentation

Start here

The gateway is OpenAI-compatible and currently serves one live model, LFM3-1.7B. The worker app runs on Windows and contributes capacity only while its local limits allow it.

API

• Quickstart — send a first request.
• Authentication — bearer keys and rate limits.
• Chat completions — request and response schema.
• Streaming — server-sent events and the final sentinel.
• Models — the active catalogue.
• Errors — status codes and error codes.

Worker

• Install the worker — requirements and enrollment.
• Worker limits — local policy and enforcement.
• Troubleshooting — heartbeat and capacity errors.

First request

Set EASYAI_API_KEY in your environment, then send a chat completion. The examples use https://<your-domain> as the gateway origin; use the origin from your invite.

curl https://<your-domain>/v1/chat/completions \
  -H "Authorization: Bearer $EASYAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "lfm3",
    "messages": [
      { "role": "user", "content": "Say hello in one sentence." }
    ]
  }'

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://<your-domain>/v1",
    api_key=os.environ["EASYAI_API_KEY"],
)

completion = client.chat.completions.create(
    model="lfm3",
    messages=[{"role": "user", "content": "Say hello in one sentence."}],
)

print(completion.choices[0].message.content)

const response = await fetch("https://<your-domain>/v1/chat/completions", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.EASYAI_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "lfm3",
    messages: [{ role: "user", content: "Say hello in one sentence." }],
  }),
});

const data = await response.json();
console.log(data.choices[0].message.content);

What is live

• LFM3-1.7B is the only live model; the model key is lfm3.
• Qwen3 models are planned, not live and not callable.
• The Windows worker app is the supported contributor client. macOS and Linux workers are not supported.
• Credits are not live yet. Terms are set per invite while the network grows.
• Pipeline parallelism is in progress and not available to contributors.
• There is no public signup, login, key creation, or installer download.

When no eligible worker is online, requests fail with 503 NO_CAPACITY. An idle network is degraded, not successful.