Migrating from OpenAI

Switch from OpenAI to llm.kiwi in minutes

llm.kiwi is designed as a drop-in replacement for the OpenAI API. Migrating requires minimal code changes—often just updating the base URL.

Quick Migration

Python

Before (OpenAI)

from openai import OpenAI

client = OpenAI(
    api_key="sk-..."
)

After (llm.kiwi)

from openai import OpenAI

client = OpenAI(
    base_url="https://api.llm.kiwi/v1",  # Add this line
    api_key="your-llm-kiwi-key"           # Use llm.kiwi key
)

Node.js

Before (OpenAI)

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: "sk-..."
});

After (llm.kiwi)

import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://api.llm.kiwi/v1',  // Add this line
  apiKey: process.env.LLM_KIWI_API_KEY  // Use llm.kiwi key
});

Model Mapping

Map OpenAI models to llm.kiwi equivalents:

OpenAI Model	llm.kiwi Equivalent	Notes
gpt-4	gpt-oss-20b	High capability open variant
gpt-4-turbo	gpt-oss-20b	Balanced performance
gpt-3.5-turbo	default	Fast routing
gpt-4o-mini	default	Model-lite routing

Use llm.kiwi’s default model to let our router automatically select the best model for each request.

API Compatibility

llm.kiwi supports most OpenAI endpoints:

Endpoint	Status
/v1/chat/completions	✅ Full support
/v1/images/generations	✅ Full support
/v1/audio/transcriptions	✅ Full support
/v1/models	✅ Full support
/v1/embeddings	⚠️ Coming soon
/v1/assistants	⚠️ Coming soon

Feature Compatibility

Feature	Supported
Streaming	✅
JSON Mode	✅ (Pro)
Function Calling	✅ (Pro)
Vision (image input)	✅ (Pro)
Temperature/Top P	✅
Max Tokens	✅

Environment Variables

Update your environment configuration:

Before

OPENAI_API_KEY=sk-...
OPENAI_BASE_URL=https://api.openai.com/v1

After

LLM_KIWI_API_KEY=your-key-here
LLM_KIWI_BASE_URL=https://api.llm.kiwi/v1

Framework-Specific Guides

LangChain

Python

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    base_url="https://api.llm.kiwi/v1",
    api_key="your-llm-kiwi-key",
    model="gpt-oss-20b"
)

LlamaIndex

Python

from llama_index.llms.openai import OpenAI

llm = OpenAI(
    api_base="https://api.llm.kiwi/v1",
    api_key="your-llm-kiwi-key",
    model="gpt-oss-20b"
)

Vercel AI SDK

TypeScript

import { createOpenAI } from '@ai-sdk/openai';

const llmkiwi = createOpenAI({
  baseURL: 'https://api.llm.kiwi/v1',
  apiKey: process.env.LLM_KIWI_API_KEY
});

Testing Your Migration

Verify your migration works correctly:

Python

# Test basic functionality
response = client.chat.completions.create(
    model="default",
    messages=[{"role": "user", "content": "Hello, is this working?"}]
)
print("✅ Basic request:", response.choices[0].message.content[:50])

# Test streaming
stream = client.chat.completions.create(
    model="default",
    messages=[{"role": "user", "content": "Count to 3"}],
    stream=True
)
print("✅ Streaming:", end=" ")
for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="")
print()

print("🎉 Migration successful!")

Common Issues

Model names don't work

OpenAI model names don’t work directly. Use llm.kiwi models:

• gpt-4 → gpt-oss-20b (or other Pro models)
• gpt-3.5-turbo → default

Missing features

Some newer OpenAI features may not be available yet. Check our API Reference for current support.

Rate limits differ

llm.kiwi has different rate limits. Review Rate Limits for your tier.

Need Help?

Contact support if you encounter migration issues.