AI API docs

Introduction

One key for Claude, GPT, Gemini, DeepSeek, Grok, Qwen, GLM and Kimi, through the OpenAI and the Anthropic API.

Most AI apps, SDKs and coding tools let you set a base URL and an API key. Put ours in those two places and the tool talks to every model we carry, with no code to rewrite. A token pack gives you one key with that many tokens; every request takes tokens from it at the model's rate.

ListingLab is an independent reseller, not affiliated with Anthropic, OpenAI, Google or the other model makers. Model names belong to their owners.

Quick start

From buying a pack to your first answer, in five steps.

1. Buy a token pack

   Pick one on the AI API page and pay by DuitNow QR or from your balance. Packs start at RM 7.90.

2. Get your key

   It is made the moment you pay and sent to your Telegram chat. It is also on your order page and on My API keys, behind Show key.

3. Check the key works

   Listing the models uses no tokens, so it is the safest first call. A list of model ids means the key is good.

   Terminal

   curl https://store.listinglab.my/v1/models \
     -H "Authorization: Bearer YOUR_KEY"

4. Ask a model something

   This uses a few tokens. Any model id from the list works; see Models & token rates.

   Terminal

   curl https://store.listinglab.my/v1/chat/completions \
     -H "Authorization: Bearer YOUR_KEY" \
     -H "Content-Type: application/json" \
     -d '{"model": "claude-sonnet-4-6", "messages": [{"role": "user", "content": "Hello"}]}'

5. Plug it into your tool

   Set the base URL and the key in Claude Code, Cursor, Cline, OpenCode or your own code.

Claude Code

Claude Code reads its address and key from environment variables, so two settings point it at us.

1. Install it

   Claude Code needs Node.js 18 or newer.

   Terminal

   npm install -g @anthropic-ai/claude-code

2. Start it on macOS or Linux

   Set the two variables, then run claude in your project folder. Add the two export lines to ~/.zshrc or ~/.bashrc to keep them.

   Terminal

   export ANTHROPIC_BASE_URL="https://store.listinglab.my"
   export ANTHROPIC_API_KEY="YOUR_KEY"
   claude

3. Start it on Windows (PowerShell)

   $env: sets them for this window; use setx to keep them.

   PowerShell

   $env:ANTHROPIC_BASE_URL = "https://store.listinglab.my"
   $env:ANTHROPIC_API_KEY = "YOUR_KEY"
   claude

4. Or set it once in settings.json

   Claude Code reads ~/.claude/settings.json when it starts, so nothing needs exporting.

   ~/.claude/settings.json

   {
     "env": {
       "ANTHROPIC_BASE_URL": "https://store.listinglab.my",
       "ANTHROPIC_API_KEY": "YOUR_KEY"
     },
     "model": "claude-sonnet-4-6"
   }

5. Pick a model

   Type /model inside Claude Code to switch, or set "model" above. Claude models suit it best.

   Claude Code

   /model claude-sonnet-4-6

Cursor

Cursor takes a custom OpenAI key and base URL in its model settings.

1. Open the model settings

   Cursor Settings, then Models.

2. Enter the key and base URL

   Under API Keys, paste your key in OpenAI API Key, turn on Override OpenAI Base URL and enter https://store.listinglab.my/v1. Then press Verify.

3. Add the models you want

   Add each model by its exact id, such as claude-sonnet-4-6, and turn it on. Pick it in the chat.

Cline / Roo Code

Both VS Code extensions take an OpenAI-compatible provider, with the same three fields.

1. Choose the provider

   Open the extension’s settings and set API Provider to OpenAI Compatible. The Anthropic provider has no base URL field, so it cannot reach us.

2. Fill in the connection

   The key goes in API Key; it is sent as a bearer token.

   Settings

   Base URL:  https://store.listinglab.my/v1
   API Key:   YOUR_KEY
   Model ID:  claude-sonnet-4-6

Kilo Code

Kilo Code connects through a custom OpenAI-compatible provider.

1. Add a provider

   Open Settings, add a custom provider, and choose the OpenAI Compatible type.

2. Name it and connect it

   The id is for Kilo; the display name is what the model picker shows.

   Settings

   Provider ID:   listinglab
   Display name:  ListingLab
   Base URL:      https://store.listinglab.my/v1
   API Key:       YOUR_KEY

3. Add models

   Add each model id exactly as the model list shows it, then pick one.

OpenCode

OpenCode keeps the key in its own credential store, so it never lands in your project.

1. Install it
   Terminal

   npm install -g opencode-ai

2. Add the provider

   In ~/.config/opencode/opencode.json (every project) or an opencode.json in the project. There is no key in this file on purpose.

   opencode.json

   {
     "$schema": "https://opencode.ai/config.json",
     "provider": {
       "listinglab": {
         "npm": "@ai-sdk/openai-compatible",
         "name": "ListingLab",
         "options": {
           "baseURL": "https://store.listinglab.my/v1"
         },
         "models": {
           "claude-sonnet-4-6": {
             "name": "claude-sonnet-4-6"
           }
         }
       }
     }
   }

3. Store the key

   Run OpenCode, type /connect, choose listinglab and paste your key. Then /models lists the models.

   Terminal

   opencode
   /connect

Python

The official OpenAI and Anthropic SDKs both work: change the base URL and the key.

OpenAI SDK

pip install openai

from openai import OpenAI

client = OpenAI(
    base_url="https://store.listinglab.my/v1",
    api_key="YOUR_KEY",
)

reply = client.chat.completions.create(
    model="claude-sonnet-4-6",
    messages=[{"role": "user", "content": "Hello"}],
)
print(reply.choices[0].message.content)

Streaming

The words arrive as they are written. A streamed answer uses the same tokens as a whole one.

stream = client.chat.completions.create(
    model="claude-sonnet-4-6",
    messages=[{"role": "user", "content": "Count to five"}],
    stream=True,
)
for chunk in stream:
    if chunk.choices and chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

Anthropic SDK

pip install anthropic

import anthropic

client = anthropic.Anthropic(
    base_url="https://store.listinglab.my",
    api_key="YOUR_KEY",
)

message = client.messages.create(
    model="claude-sonnet-4-6",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello"}],
)
print(message.content[0].text)

Node.js

The official OpenAI package, with our base URL.

npm install openai

import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://store.listinglab.my/v1',
  apiKey: process.env.LISTINGLAB_AI_KEY,
});

const reply = await client.chat.completions.create({
  model: 'claude-sonnet-4-6',
  messages: [{ role: 'user', content: 'Hello' }],
});
console.log(reply.choices[0].message.content);

Streaming

const stream = await client.chat.completions.create({
  model: 'claude-sonnet-4-6',
  messages: [{ role: 'user', content: 'Count to five' }],
  stream: true,
});
for await (const chunk of stream) process.stdout.write(chunk.choices[0]?.delta?.content ?? '');

cURL / HTTP

Both formats over plain HTTP.

OpenAI format

The key goes in Authorization: Bearer.

curl https://store.listinglab.my/v1/chat/completions \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-sonnet-4-6", "messages": [{"role": "user", "content": "Hello"}]}'

Anthropic format

The key goes in x-api-key, with the anthropic-version header.

curl https://store.listinglab.my/v1/messages \
  -H "x-api-key: YOUR_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-sonnet-4-6", "max_tokens": 1024, "messages": [{"role": "user", "content": "Hello"}]}'

Streaming

Add "stream": true; the answer arrives as server-sent events and ends with data: [DONE].

curl -N https://store.listinglab.my/v1/chat/completions \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-sonnet-4-6", "stream": true, "messages": [{"role": "user", "content": "Count to five"}]}'

OpenAI-compatible mode

Any OpenAI client works against https://store.listinglab.my/v1.

Requests and answers follow the OpenAI chat completions shape: system messages, temperature, top_p, max_tokens, stop and streaming. Tool calling (tools and tool_choice) is passed to the model as it is, so it works where the model supports it. Pictures go in as image_url parts, to models that can see.

Use the Anthropic format (/v1/messages) for tools built around it, such as Claude Code, and the OpenAI format for nearly everything else. Both cost the same tokens.

Models & token rates

Each model takes tokens from your key at its own rate.

A request uses the tokens it reads (your message, and the whole chat so far) plus the tokens it writes, times the model's rate. At 1.8×, a request of 2,000 tokens takes 3,600 from your key; at 0.7× it takes 1,400. Cached input, text the model has seen moments before, is charged at the lower cached rate. Requests that fail use nothing.

46 of 49 models answered at the last check, 2026-10-14 08:41:05. Every model is asked for one word every 30 minutes.

Use the model id exactly as shown. GET /v1/models with your key gives the same list and uses no tokens.

Your key & usage

Where your key is, what is left on it, and keeping it safe.

• Finding it. The key comes to your Telegram chat when you pay, and is on your order page. On My API keys, Show key reveals it again any time.
• Tokens left. My API keys shows each key's tokens left, tokens used and requests, with every request listed by model, tokens and time. It updates within a minute or two of each request.
• One pack, one key. Each pack you buy is a separate key with its own tokens. Keep one per project, or buy another pack when a key runs low.
• Expiry. No expiry date is set on the keys; the tokens stay until you use them.
• Keep it secret. Anyone with the key can spend its tokens. Never put it in a web page, an app you hand out, or a public repository; read it from an environment variable. In the Playground the key never leaves our server.

Errors

What a refused request means, and what to do.