Stormfire — Python SDK

Drop-in replacement for the OpenAI Python SDK that routes your API calls through Stormfire — pay AI APIs in USDT, no Stripe, no KYC.

Install

pip install stormfire

Quick start

from stormfire import Stormfire

client = Stormfire(api_key="sk-...")  # get yours at https://stormfire.io

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

That's it. Same interface as openai.OpenAI, paid in USDT.

Why?

Stormfire is an AI API gateway. You pay us in USDT, we pay OpenAI / Anthropic / Google / DeepSeek on your behalf. No Stripe in the loop, no KYC, works in 30+ countries that traditional billing rejects.

This package is a 30-line wrapper around the official openai SDK that:

• Sets the base URL to https://api.stormfire.io/v1
• Reads STORMFIRE_API_KEY (falls back to OPENAI_API_KEY)
• Exports both sync (Stormfire) and async (AsyncStormfire) clients

Everything else is identical. All openai features work — streaming, function calling, vision, structured outputs, the whole surface area.

Migrate an existing project

If you already have OpenAI SDK code, you have two options:

Option 1 — Use this package (recommended)

# before
from openai import OpenAI
client = OpenAI()

# after
from stormfire import Stormfire
client = Stormfire()  # picks up STORMFIRE_API_KEY or OPENAI_API_KEY

Option 2 — Just change the base URL on the official SDK

from openai import OpenAI

client = OpenAI(
    api_key="sk-...",
    base_url="https://api.stormfire.io/v1",
)

Both work. Option 1 is shorter and survives upstream renames better.

Async usage

import asyncio
from stormfire import AsyncStormfire

async def main():
    client = AsyncStormfire()
    response = await client.chat.completions.create(
        model="claude-3-5-sonnet",
        messages=[{"role": "user", "content": "Write a haiku"}],
    )
    print(response.choices[0].message.content)

asyncio.run(main())

Streaming

stream = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Count to 10"}],
    stream=True,
)
for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="")

Models

Stormfire supports 30+ models including:

• OpenAI: gpt-4o, gpt-5, gpt-4o-mini
• Anthropic: claude-3-5-sonnet, claude-3-7-sonnet, claude-3-5-haiku
• Google: gemini-2.0-pro, gemini-2.0-flash
• DeepSeek: deepseek-v3, deepseek-v4
• Mistral: mistral-large, mistral-small
• Open source: llama-3.1-405b, qwen-2.5-72b

Full list and current pricing: https://stormfire.io/pricing.

LangChain / LlamaIndex / etc.

Anything that accepts a base_url works. Example with LangChain:

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    api_key="sk-stormfire-key",
    base_url="https://api.stormfire.io/v1",
    model="gpt-4o",
)

Development

git clone https://github.com/stormfire-io/stormfire-python
cd stormfire-python
pip install -e ".[dev]"
pytest

License

MIT — see LICENSE.