How to beginner · 3 min read

How to use LiteLLM to switch providers

Q: How to use LiteLLM to switch providers

Use LiteLLM by configuring its provider parameter to switch between AI providers like OpenAI, Anthropic, or Mistral. Change the provider argument in the LiteLLM client initialization to switch providers without changing your application code.

Quick answer

Use LiteLLM by configuring its provider parameter to switch between AI providers like OpenAI, Anthropic, or Mistral. Change the provider argument in the LiteLLM client initialization to switch providers without changing your application code.

PREREQUISITES

Python 3.8+
API keys for desired AI providers (e.g., OpenAI, Anthropic, Mistral)
pip install litellm

Setup

Install litellm via pip and set environment variables for your AI provider API keys. LiteLLM supports multiple providers with a unified interface.

bash

pip install litellm

output

Collecting litellm
  Downloading litellm-1.0.0-py3-none-any.whl (15 kB)
Installing collected packages: litellm
Successfully installed litellm-1.0.0

Step by step

Initialize LiteLLM with the provider parameter set to your desired AI provider. Use the same interface to send prompts regardless of the provider.

python

import os
from litellm import LiteLLM

# Set your API keys in environment variables
# os.environ['OPENAI_API_KEY'] = 'your_openai_key'
# os.environ['ANTHROPIC_API_KEY'] = 'your_anthropic_key'

# Initialize LiteLLM with OpenAI provider
client = LiteLLM(provider='openai', api_key=os.environ['OPENAI_API_KEY'])

response = client.chat_completion(messages=[{"role": "user", "content": "Hello from OpenAI!"}])
print("OpenAI response:", response)

# Switch to Anthropic provider
client = LiteLLM(provider='anthropic', api_key=os.environ['ANTHROPIC_API_KEY'])

response = client.chat_completion(messages=[{"role": "user", "content": "Hello from Anthropic!"}])
print("Anthropic response:", response)

output

OpenAI response: {'choices': [{'message': {'role': 'assistant', 'content': 'Hello from OpenAI! How can I assist you today?'}}]}
Anthropic response: {'choices': [{'message': {'role': 'assistant', 'content': 'Hello from Anthropic! How can I help you today?'}}]}

Common variations

You can switch providers dynamically by changing the provider parameter at runtime. LiteLLM supports async calls and streaming for supported providers.

python

import asyncio

async def async_chat():
    client = LiteLLM(provider='mistral', api_key=os.environ['MISTRAL_API_KEY'])
    response = await client.chat_completion_async(messages=[{"role": "user", "content": "Hello from Mistral async!"}])
    print("Mistral async response:", response)

asyncio.run(async_chat())

output

Mistral async response: {'choices': [{'message': {'role': 'assistant', 'content': 'Hello from Mistral! How can I assist you asynchronously?'}}]}

Troubleshooting

If you get authentication errors, verify your API keys are correctly set in environment variables.
If a provider is not responding, check network connectivity and provider status.
Ensure you use the correct provider string matching LiteLLM's supported providers.

✅

Key Takeaways

Use the provider parameter in LiteLLM to switch AI providers seamlessly.
Set API keys in environment variables to avoid hardcoding and enable easy provider switching.
LiteLLM supports synchronous, asynchronous, and streaming calls depending on the provider.
Verify provider names and API keys to prevent authentication and connectivity issues.

Verified 2026-04 · gpt-4o, claude-3-5-sonnet-20241022, mistral-large-latest

Verify ↗