Quickstart: base URL, API key, first request

Send your first request

The API is OpenAI-compatible. Point your existing code at this base URL and keep everything else.

Anthropic-style clients are the one exception — see Base URL below.

Try it with curl

# replace YOUR_KEY with the key from your dashboard
curl https://apiclan.us/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_KEY" \
  -d '{
    "model": "gpt-5.6-sol",
    "messages": [{"role": "user", "content": "Say hello in five words."}]
  }'

A JSON response means you are done. Anything else — read Something went wrong, it covers every case we have seen.

Get an API key

• Sign in, open API Keys in the dashboard, and create a key.
• A key belongs to one group. The group decides which models it can reach and at what rate, so a key created for one plan will not work on another.
• Your key stays visible in the console, so you can copy it again later. That makes your account login the real secret — and each key can be locked to the IP addresses you allow.

Base URL

Every request has to arrive at a path starting with /v1. Whether you put /v1 in the base URL depends on your client, because some of them append it for you and some do not.

Code samples

Python — OpenAI SDK

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_KEY",
    base_url="https://apiclan.us/v1",   # the only line that changes
)

resp = client.chat.completions.create(
    model="gpt-5.6-sol",
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

Node — OpenAI SDK

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "YOUR_KEY",
  baseURL: "https://apiclan.us/v1",
});

const r = await client.chat.completions.create({
  model: "claude-opus-5",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(r.choices[0].message.content);

Streaming

Streaming works exactly as it does upstream — pass stream: true. We do not buffer responses, so tokens reach you as they are produced.

Desktop & CLI tools

Claude Code

Set two environment variables. Note the base URL carries no /v1 — Claude Code adds /v1/messages itself.

export ANTHROPIC_BASE_URL="https://apiclan.us"
export ANTHROPIC_AUTH_TOKEN="YOUR_KEY"

Codex

Add a custom provider in ~/.codex/config.toml and select it:

model_provider = "apiclan"
model = "gpt-5.6-sol"

[model_providers.apiclan]
name = "APICLAN"
base_url = "https://apiclan.us/v1"
env_key = "APICLAN_API_KEY"

Then put your key in the APICLAN_API_KEY environment variable.

Cherry Studio, Chatbox, LobeChat and similar

Add a provider of type OpenAI, paste https://apiclan.us/v1 as the API host and your key as the API key, then add the model names you want from the list below. These apps do not discover models automatically.

LM Studio

LM Studio can talk to a remote OpenAI-compatible endpoint instead of a local model. Open Settings → Model Provider, add a custom provider, then fill in:

Keep the /v1 in the base URL. LM Studio sends requests to the path exactly as you type it and does not append /v1 for you.

Cursor, Cline and Roo Code

All three accept a custom OpenAI-compatible endpoint. The base URL is https://apiclan.us/v1 in each of them.

• Cursor — Settings → Models → enable Override OpenAI Base URL, paste the base URL and key, then add the model IDs you want by hand.
• Cline and Roo Code — pick OpenAI Compatible as the provider, then fill in base URL, key and model ID.

Using Claude Fable 5.1

claude-fable-5-1 is Anthropic’s long-horizon model — built for agent loops that run for hours, multi-file refactors and multistep research. It is available here on the Claude CC Max group. Three things about it will break code written for other Claude models:

• Forced tool use is gone. tool_choice set to {"type":"any"} or {"type":"tool"} returns a 400. Thinking is always on for this model, and a forced call would skip it. Keep tool_choice: {"type":"auto"} and say in the prompt when the tool applies.
• Cache reads cost a quarter of what they do elsewhere. $0.25 per million tokens — 0.025× base input, against 0.1× on every other Claude model. In a long agent loop most of your input is cache reads, so the model that looks expensive per token can end up cheaper per session.
• One million tokens of context, at standard per-token pricing across the whole window, with 128k max output.

It also declines non-default temperature, top_p and top_k with a 400, and prefilling the assistant response is not supported. Those apply to claude-fable-5 too.

Generating images

Image generation draws on the same balance, through the standard OpenAI-compatible endpoint — with a key from an image group (see the note below):

curl https://apiclan.us/v1/images/generations \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-image-2.5","prompt":"a red maple leaf on white paper",
       "size":"1024x1024","n":1,"response_format":"b64_json"}'

Every image comes back at about 1.6 megapixels and is billed at the 1K price. Keep the long edge of size at 1024 and set the shape you want: 1024x1024 (1:1), 1024x576 (16:9), 576x1024 (9:16). A larger size does not give you a larger image — it only moves the request into a higher billing tier. Models: gpt-image-2.5, gpt-image-2.5-flare, gpt-image-2.5-sunburst and gpt-image-2. There is also a page in the dashboard — AI Image in the sidebar — if you would rather type a prompt than write a request.

Models

Use these exact strings as the model value. Current prices are on the homepage — we keep them there rather than here so you always see live numbers.

Per-model deep dives, each built from real invoices on this account — what a model actually costs on a working day, how much of that is cache reads, and when a cheaper model in the same family does the same job: GPT-6 Astra, GPT-5.6 Sol, GPT-5.6 Terra, Claude Fable 5.1, Claude Opus 5, Claude Sonnet 5, GPT Image 2.5.

You can also fetch the list your key can reach:

curl https://apiclan.us/v1/models -H "Authorization: Bearer YOUR_KEY"

Credits & top-ups

• Pay as you go. No subscription, no monthly minimum, no commitment.
• Top up with USDT — 1 USDT gives you 2 credits of API balance.
• Prices are per million tokens, charged separately for input and output.

Redeem codes

Fixed amounts are also sold as redeem codes at /buy. No account is needed to buy one, the code works on any APICLAN account, and you can pass it on. Redeem it in the dashboard under Redeem code.

Something went wrong