Quickstart

Everything is OpenAI-compatible. Point your existing SDK at our base URL.

1. Get a key

Sign in to the console, top up, and create an API token.

2. Call the API

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.apilili.com/v1",
  apiKey: "sk-your-token",
});

const res = await client.chat.completions.create({
  model: "deepseek-chat",
  messages: [{ role: "user", content: "Hello!" }],
});
console.log(res.choices[0].message.content);

3. Available models

Run GET /v1/models with your key for the live list. Currently: DeepSeek V4 Flash/Pro, GLM-5.3 Flash, GLM-4.7, Qwen-Plus, Kimi K3.

Get your API key

1. Register in the console (email + password)
2. Top up (redemption code or pay page)
3. Create a token in the Tokens page — that is your API key (sk-...)
4. Use it in any OpenAI-compatible client. Quota and limits are visible in the console.

Use from your local apps

This is an API service — bring your own app. No web chat required.

• Python / Node OpenAI SDK: set base_url to https://api.apilili.com/v1 and your API key
• curl: POST /v1/chat/completions with your key (see example above)
• LangChain / LlamaIndex: use the OpenAI-compatible ChatOpenAI with our base URL
• Dify / Cherry Studio / NextChat / Cline: add a custom model provider with our base URL + key
• Local agents (OpenClaw, OpenCode, Cursor): configure an OpenAI-compatible endpoint

Anthropic-compatible endpoint

Select models also expose an Anthropic-format endpoint. Contact support for details.

Rate limits & quotas

• Your token shows its remaining quota in the console.
• Default rate limit: 60 requests/min per token (adjustable on Business plans).
• Max request body: 20 MB.

Errors