Documentation · APIHub

Quickstart

Our API is wire-compatible with OpenAI's. Any client, SDK or tool that lets you set a custom base URL will work without modification.

1Create a key on the API Keys page in your console. Copy it immediately — the full value is shown once and cannot be retrieved later.
2Change the base URL in your code to the gateway address below.
3Replace your official API key with the one you just created, and leave every other parameter as it is.

Endpoints

Pick the base URL matching your SDK's protocol — the client appends the endpoint path to it. The same API key works for both.

OpenAI / GPT protocol (recommended)

# Base URL
https://apihub.ltd/v1

# Chat Completions
https://apihub.ltd/v1/chat/completions

# Responses
https://apihub.ltd/v1/responses

Anthropic / Claude protocol

# Base URL
https://apihub.ltd

# Messages
https://apihub.ltd/v1/messages

The OpenAI base URL already includes /v1, and clients append /chat/completions or /responses to it; the Anthropic base URL is the root, and clients append /v1/messages. If a duplicated /v1 causes a 404, adjust the base URL accordingly.

Authentication

Every request carries your API key in an HTTP header. Use whichever form matches your protocol.

OpenAI

Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

Anthropic

x-api-key: YOUR_API_KEY
anthropic-version: 2023-06-01
Content-Type: application/json

Streaming

Set stream to true to receive SSE chunks. The event format matches the official one exactly and terminates with data: [DONE].

curl https://apihub.ltd/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-5",
    "messages": [{"role": "user", "content": "Hello!"}],
    "stream": true
  }'

# data: {"choices":[{"delta":{"content":"He"}}]}
# data: {"choices":[{"delta":{"content":"llo"}}]}
# data: [DONE]

Error codes

Error responses mirror the official shape, so your existing error handling keeps working.

Status	Meaning	What to do
401	Key is invalid, disabled or expired	Check the key was copied in full, or verify its status in the console.
402	Insufficient balance	Add credit, or check whether the key hit its own spend limit.
403	Key may not access the requested model	Review the model allowlist configured on that key.
404	Model not found or not enabled	Check the model parameter against the pricing page.
429	Rate limit exceeded	Lower concurrency or request rate, and retry with exponential backoff.
500 / 502	Upstream provider error	We fail over automatically; contact support if it persists.

Third-party clients

These all work out of the box — just set a custom base URL and API key in their settings.

Desktop and web clients such as Cherry Studio, ChatBox and LobeChat
Coding assistants such as Cursor, Continue and Cline
Workflow and RAG platforms such as Dify, n8n and FastGPT
Any in-house application built on the official OpenAI SDK

Can't find what you need?

Contact support