🚅 LiteLLM

Call all LLM APIs using the OpenAI format [Bedrock, Huggingface, Cohere, TogetherAI, Azure, OpenAI, etc.]

Evaluate LLMs → OpenAI-Compatible Server

LiteLLM manages

• Translating inputs to the provider's completion and embedding endpoints
• Guarantees consistent output, text responses will always be available at ['choices'][0]['message']['content']
• Exception mapping - common exceptions across providers are mapped to the OpenAI exception types.
• Load-balance across multiple deployments (e.g. Azure/OpenAI) - Router

Usage (Docs)

Important

LiteLLM v1.0.0 is now requires openai>=1.0.0. Migration guide here

pip install litellm

from litellm import completion
import os

## set ENV variables 
os.environ["OPENAI_API_KEY"] = "your-openai-key" 
os.environ["COHERE_API_KEY"] = "your-cohere-key" 

messages = [{ "content": "Hello, how are you?","role": "user"}]

# openai call
response = completion(model="gpt-3.5-turbo", messages=messages)

# cohere call
response = completion(model="command-nightly", messages=messages)
print(response)

Streaming (Docs)

liteLLM supports streaming the model response back, pass stream=True to get a streaming iterator in response.
Streaming is supported for all models (Bedrock, Huggingface, TogetherAI, Azure, OpenAI, etc.)

from litellm import completion
response = completion(model="gpt-3.5-turbo", messages=messages, stream=True)
for part in response:
    print(part.choices[0].delta.content or "")

# claude 2
response = completion('claude-2', messages, stream=True)
for part in response:
    print(part.choices[0].delta.content or "")

Router - load balancing(Docs)

LiteLLM allows you to load balance between multiple deployments (Azure, OpenAI). It picks the deployment which is below rate-limit and has the least amount of tokens used.

from litellm import Router

model_list = [{ # list of model deployments 
    "model_name": "gpt-3.5-turbo", # model alias 
    "litellm_params": { # params for litellm completion/embedding call 
        "model": "azure/chatgpt-v-2", # actual model name
        "api_key": os.getenv("AZURE_API_KEY"),
        "api_version": os.getenv("AZURE_API_VERSION"),
        "api_base": os.getenv("AZURE_API_BASE")
    }
}, {
    "model_name": "gpt-3.5-turbo", 
    "litellm_params": { # params for litellm completion/embedding call 
        "model": "azure/chatgpt-functioncalling", 
        "api_key": os.getenv("AZURE_API_KEY"),
        "api_version": os.getenv("AZURE_API_VERSION"),
        "api_base": os.getenv("AZURE_API_BASE")
    }
}, {
    "model_name": "gpt-3.5-turbo", 
    "litellm_params": { # params for litellm completion/embedding call 
        "model": "gpt-3.5-turbo", 
        "api_key": os.getenv("OPENAI_API_KEY"),
    }
}]

router = Router(model_list=model_list)

# openai.ChatCompletion.create replacement
response = router.completion(model="gpt-3.5-turbo", 
                messages=[{"role": "user", "content": "Hey, how's it going?"}])

print(response)

OpenAI Proxy - (Docs)

If you want to use non-openai models in an openai code base, you can use litellm proxy. Create a server to call 100+ LLMs (Huggingface/Bedrock/TogetherAI/etc) in the OpenAI ChatCompletions & Completions format

Step 1: Start litellm proxy

$ litellm --model huggingface/bigcode/starcoder

#INFO: Proxy running on http://0.0.0.0:8000

Step 2: Replace openai base

import openai # openai v1.0.0+
client = openai.OpenAI(api_key="anything",base_url="http://0.0.0.0:8000") # set proxy to base_url
# request sent to model set on litellm proxy, `litellm --model`
response = client.chat.completions.create(model="gpt-3.5-turbo", messages = [
    {
        "role": "user",
        "content": "this is a test request, write a short poem"
    }
])

print(response)

Logging Observability (Docs)

LiteLLM exposes pre defined callbacks to send data to Langfuse, LLMonitor, Helicone, Promptlayer, Traceloop, Slack

from litellm import completion

## set env variables for logging tools
os.environ["LANGFUSE_PUBLIC_KEY"] = ""
os.environ["LANGFUSE_SECRET_KEY"] = ""
os.environ["LLMONITOR_APP_ID"] = "your-llmonitor-app-id"

os.environ["OPENAI_API_KEY"]

# set callbacks
litellm.success_callback = ["langfuse", "llmonitor"] # log input/output to langfuse, llmonitor, supabase

#openai call
response = completion(model="gpt-3.5-turbo", messages=[{"role": "user", "content": "Hi 👋 - i'm openai"}])

Supported Provider (Docs)

Provider	Completion	Streaming	Async Completion	Async Streaming
openai	✅	✅	✅	✅
azure	✅	✅	✅	✅
aws - sagemaker	✅	✅	✅	✅
aws - bedrock	✅	✅	✅	✅
cohere	✅	✅	✅	✅
anthropic	✅	✅	✅	✅
huggingface	✅	✅	✅	✅
replicate	✅	✅	✅	✅
together_ai	✅	✅	✅	✅
openrouter	✅	✅	✅	✅
google - vertex_ai	✅	✅	✅	✅
google - palm	✅	✅	✅	✅
ai21	✅	✅	✅	✅
baseten	✅	✅	✅	✅
vllm	✅	✅	✅	✅
nlp_cloud	✅	✅	✅	✅
aleph alpha	✅	✅	✅	✅
petals	✅	✅	✅	✅
ollama	✅	✅	✅	✅
deepinfra	✅	✅	✅	✅
perplexity-ai	✅	✅	✅	✅
anyscale	✅	✅	✅	✅

Read the Docs

Contributing

To contribute: Clone the repo locally -> Make a change -> Submit a PR with the change.

Here's how to modify the repo locally: Step 1: Clone the repo

git clone https://github.com/BerriAI/litellm.git

Step 2: Navigate into the project, and install dependencies:

cd litellm
poetry install

Step 3: Test your change:

cd litellm/tests # pwd: Documents/litellm/litellm/tests
pytest .

Step 4: Submit a PR with your changes! 🚀

• push your fork to your GitHub repo
• submit a PR from there

Support / talk with founders

• Schedule Demo 👋
• Community Discord 💭
• Our numbers 📞 +1 (770) 8783-106 / ‭+1 (412) 618-6238‬
• Our emails ✉️ [email protected] / [email protected]

Why did we build this

• Need for simplicity: Our code started to get extremely complicated managing & translating calls between Azure, OpenAI and Cohere.

Contributors