LiteLLM - Getting Started
https://github.com/BerriAI/litellm
Call 100+ LLMs using the same Input/Output Format​
- Translate inputs to provider's
completion
,embedding
, andimage_generation
endpoints - Consistent output, text responses will always be available at
['choices'][0]['message']['content']
- Retry/fallback logic across multiple deployments (e.g. Azure/OpenAI) - Router
- Track spend & set budgets per project OpenAI Proxy Server
How to use LiteLLM​
You can use litellm through either:
- OpenAI proxy Server - Server to call 100+ LLMs, load balance, cost tracking across projects
- LiteLLM python SDK - Python Client to call 100+ LLMs, load balance, cost tracking
LiteLLM Python SDK​
Basic usage​
pip install litellm
- OpenAI
- Anthropic
- VertexAI
- HuggingFace
- Azure OpenAI
- Ollama
- Openrouter
from litellm import completion
import os
## set ENV variables
os.environ["OPENAI_API_KEY"] = "your-api-key"
response = completion(
model="gpt-3.5-turbo",
messages=[{ "content": "Hello, how are you?","role": "user"}]
)
from litellm import completion
import os
## set ENV variables
os.environ["ANTHROPIC_API_KEY"] = "your-api-key"
response = completion(
model="claude-2",
messages=[{ "content": "Hello, how are you?","role": "user"}]
)
from litellm import completion
import os
# auth: run 'gcloud auth application-default'
os.environ["VERTEX_PROJECT"] = "hardy-device-386718"
os.environ["VERTEX_LOCATION"] = "us-central1"
response = completion(
model="chat-bison",
messages=[{ "content": "Hello, how are you?","role": "user"}]
)
from litellm import completion
import os
os.environ["HUGGINGFACE_API_KEY"] = "huggingface_api_key"
# e.g. Call 'WizardLM/WizardCoder-Python-34B-V1.0' hosted on HF Inference endpoints
response = completion(
model="huggingface/WizardLM/WizardCoder-Python-34B-V1.0",
messages=[{ "content": "Hello, how are you?","role": "user"}],
api_base="https://my-endpoint.huggingface.cloud"
)
print(response)
from litellm import completion
import os
## set ENV variables
os.environ["AZURE_API_KEY"] = ""
os.environ["AZURE_API_BASE"] = ""
os.environ["AZURE_API_VERSION"] = ""
# azure call
response = completion(
"azure/<your_deployment_name>",
messages = [{ "content": "Hello, how are you?","role": "user"}]
)
from litellm import completion
response = completion(
model="ollama/llama2",
messages = [{ "content": "Hello, how are you?","role": "user"}],
api_base="http://localhost:11434"
)
from litellm import completion
import os
## set ENV variables
os.environ["OPENROUTER_API_KEY"] = "openrouter_api_key"
response = completion(
model="openrouter/google/palm-2-chat-bison",
messages = [{ "content": "Hello, how are you?","role": "user"}],
)
Streaming​
Set stream=True
in the completion
args.
- OpenAI
- Anthropic
- VertexAI
- HuggingFace
- Azure OpenAI
- Ollama
- Openrouter
from litellm import completion
import os
## set ENV variables
os.environ["OPENAI_API_KEY"] = "your-api-key"
response = completion(
model="gpt-3.5-turbo",
messages=[{ "content": "Hello, how are you?","role": "user"}],
stream=True,
)
from litellm import completion
import os
## set ENV variables
os.environ["ANTHROPIC_API_KEY"] = "your-api-key"
response = completion(
model="claude-2",
messages=[{ "content": "Hello, how are you?","role": "user"}],
stream=True,
)
from litellm import completion
import os
# auth: run 'gcloud auth application-default'
os.environ["VERTEX_PROJECT"] = "hardy-device-386718"
os.environ["VERTEX_LOCATION"] = "us-central1"
response = completion(
model="chat-bison",
messages=[{ "content": "Hello, how are you?","role": "user"}],
stream=True,
)
from litellm import completion
import os
os.environ["HUGGINGFACE_API_KEY"] = "huggingface_api_key"
# e.g. Call 'WizardLM/WizardCoder-Python-34B-V1.0' hosted on HF Inference endpoints
response = completion(
model="huggingface/WizardLM/WizardCoder-Python-34B-V1.0",
messages=[{ "content": "Hello, how are you?","role": "user"}],
api_base="https://my-endpoint.huggingface.cloud",
stream=True,
)
print(response)
from litellm import completion
import os
## set ENV variables
os.environ["AZURE_API_KEY"] = ""
os.environ["AZURE_API_BASE"] = ""
os.environ["AZURE_API_VERSION"] = ""
# azure call
response = completion(
"azure/<your_deployment_name>",
messages = [{ "content": "Hello, how are you?","role": "user"}],
stream=True,
)
from litellm import completion
response = completion(
model="ollama/llama2",
messages = [{ "content": "Hello, how are you?","role": "user"}],
api_base="http://localhost:11434",
stream=True,
)
from litellm import completion
import os
## set ENV variables
os.environ["OPENROUTER_API_KEY"] = "openrouter_api_key"
response = completion(
model="openrouter/google/palm-2-chat-bison",
messages = [{ "content": "Hello, how are you?","role": "user"}],
stream=True,
)
Exception handling​
LiteLLM maps exceptions across all supported providers to the OpenAI exceptions. All our exceptions inherit from OpenAI's exception types, so any error-handling you have for that, should work out of the box with LiteLLM.
from openai.error import OpenAIError
from litellm import completion
os.environ["ANTHROPIC_API_KEY"] = "bad-key"
try:
# some code
completion(model="claude-instant-1", messages=[{"role": "user", "content": "Hey, how's it going?"}])
except OpenAIError as e:
print(e)
Logging Observability - Log LLM Input/Output (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"}])
Track Costs, Usage, Latency for streaming​
Use a callback function for this - more info on custom callbacks: https://docs.litellm.ai/docs/observability/custom_callback
import litellm
# track_cost_callback
def track_cost_callback(
kwargs, # kwargs to completion
completion_response, # response from completion
start_time, end_time # start/end time
):
try:
response_cost = kwargs.get("response_cost", 0)
print("streaming response_cost", response_cost)
except:
pass
# set callback
litellm.success_callback = [track_cost_callback] # set custom callback function
# litellm.completion() call
response = completion(
model="gpt-3.5-turbo",
messages=[
{
"role": "user",
"content": "Hi 👋 - i'm openai"
}
],
stream=True
)
OpenAI Proxy​
Track spend across multiple projects/people
The proxy provides:
📖 Proxy Endpoints - Swagger Docs​
Quick Start Proxy - CLI​
pip install 'litellm[proxy]'
Step 1: Start litellm proxy​
$ litellm --model huggingface/bigcode/starcoder
#INFO: Proxy running on http://0.0.0.0:4000
Step 2: Make ChatCompletions Request to Proxy​
import openai # openai v1.0.0+
client = openai.OpenAI(api_key="anything",base_url="http://0.0.0.0:4000") # 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)