OpenAI Compatible API Guide

The OpenAI API format has become the standard interface for large language models. An "OpenAI-compatible API" accepts the same request format (/v1/chat/completions) and returns responses in the same structure—letting you swap models without changing your code.

APIMaster.ai provides an OpenAI-compatible endpoint for Claude, GPT, DeepSeek, Gemini, and more.

What Is an OpenAI-Compatible API?

Any API that implements these endpoints is "OpenAI-compatible":

Endpoint	Method	Purpose
/v1/chat/completions	POST	Generate responses
/v1/models	GET	List available models
/v1/embeddings	POST	Create embeddings

The request/response schema is identical to OpenAI's, so any tool that supports OpenAI (LangChain, LiteLLM, Dify, open-webui, etc.) works automatically.

Using APIMaster's OpenAI-Compatible Endpoint

Base URL: https://apimaster.ai/v1

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_APIMASTER_KEY",
    base_url="https://apimaster.ai/v1",
)

# Call Claude
response = client.chat.completions.create(
    model="claude-sonnet-4-6",
    messages=[{"role": "user", "content": "Hello from OpenAI-compatible API!"}],
)

# Call DeepSeek
response = client.chat.completions.create(
    model="deepseek-v4",
    messages=[{"role": "user", "content": "Hello!"}],
)

# Call GPT
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello!"}],
)

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

One client, three providers, zero code changes between them.

Integrating with Popular Frameworks

LangChain

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    model="claude-sonnet-4-6",
    openai_api_key="YOUR_APIMASTER_KEY",
    openai_api_base="https://apimaster.ai/v1",
)

response = llm.invoke("What is LangChain used for?")
print(response.content)

LiteLLM

import litellm

response = litellm.completion(
    model="openai/claude-sonnet-4-6",
    api_key="YOUR_APIMASTER_KEY",
    api_base="https://apimaster.ai/v1",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(response.choices[0].message.content)

Dify

In Dify's model provider settings:

• Provider: OpenAI-Compatible
• API Key: your APIMaster key
• API Base URL: https://apimaster.ai/v1
• Model name: claude-sonnet-4-6 or gpt-4o

Open WebUI

docker run -d \
  -e OPENAI_API_KEY="YOUR_APIMASTER_KEY" \
  -e OPENAI_API_BASE_URL="https://apimaster.ai/v1" \
  -p 3000:8080 \
  ghcr.io/open-webui/open-webui

Cursor / VS Code AI Extensions

Set Custom OpenAI Base URL to https://apimaster.ai/v1 and use your APIMaster key. Select any model from the available list.

Available Models on APIMaster OpenAI-Compatible Endpoint

Claude (Anthropic)

• claude-sonnet-4-6 — best value
• claude-opus-4-8 — most capable
• claude-haiku-4-5 — fastest, cheapest
• claude-fable-5 — creative

GPT (OpenAI)

• gpt-5 — latest flagship
• gpt-4o — multimodal
• gpt-4o-mini — budget

DeepSeek

• deepseek-v4 — strong coding, very cheap
• deepseek-r1 — reasoning

Gemini (Google)

• gemini-2.5-pro — long context

See the full list on the models page.

Model Switching at Runtime

MODELS = ["claude-sonnet-4-6", "gpt-4o", "deepseek-v4"]

def compare_models(prompt):
    client = OpenAI(api_key="YOUR_KEY", base_url="https://apimaster.ai/v1")
    for model in MODELS:
        resp = client.chat.completions.create(
            model=model,
            messages=[{"role": "user", "content": prompt}],
            max_tokens=200,
        )
        print(f"\n=== {model} ===")
        print(resp.choices[0].message.content)

compare_models("Explain recursion in one paragraph.")

Migrating from OpenAI to OpenAI-Compatible API

Change exactly two lines:

# Before
from openai import OpenAI
client = OpenAI(api_key="sk-...")

# After (APIMaster)
from openai import OpenAI
client = OpenAI(
    api_key="YOUR_APIMASTER_KEY",   # ← changed
    base_url="https://apimaster.ai/v1",         # ← added
)

Everything else—models, streaming, function calling, embeddings—stays the same.

Why APIMaster for OpenAI-Compatible API

• All major providers in one endpoint: no separate keys for each provider
• Fingerprint verified: know which model you're actually calling
• Works worldwide: no VPN required for China/Russia users
• Discounted pricing: up to 90% off on select models

Get API access → · See model pricing →