2023-08-24 22:51:02 +08:00
< h1 align = "center" >
2023-08-24 23:32:50 +08:00
🚅 LiteLLM
< / h1 >
< p align = "center" >
2024-05-25 00:19:28 +08:00
< p align = "center" >
2024-05-25 00:19:48 +08:00
< a href = "https://render.com/deploy?repo=https://github.com/BerriAI/litellm" target = "_blank" rel = "nofollow" > < img src = "https://render.com/images/deploy-to-render-button.svg" alt = "Deploy to Render" > < / a >
2024-05-25 00:07:32 +08:00
< a href = "https://railway.app/template/HLP0Ub?referralCode=jch2ME" >
< img src = "https://railway.app/button.svg" alt = "Deploy on Railway" >
< / a >
2024-05-25 00:19:28 +08:00
< / p >
2024-07-25 07:51:40 +08:00
< p align = "center" > Call all LLM APIs using the OpenAI format [Bedrock, Huggingface, VertexAI, TogetherAI, Azure, OpenAI, Groq etc.]
2023-10-08 06:57:44 +08:00
< br >
2023-08-24 23:32:50 +08:00
< / p >
2024-08-09 08:06:29 +08:00
< h4 align = "center" > < a href = "https://docs.litellm.ai/docs/simple_proxy" target = "_blank" > LiteLLM Proxy Server (LLM Gateway)< / a > | < a href = "https://docs.litellm.ai/docs/hosted" target = "_blank" > Hosted Proxy (Preview)< / a > | < a href = "https://docs.litellm.ai/docs/enterprise" target = "_blank" > Enterprise Tier< / a > < / h4 >
2023-08-24 22:55:02 +08:00
< h4 align = "center" >
2023-08-24 23:32:50 +08:00
< a href = "https://pypi.org/project/litellm/" target = "_blank" >
2023-08-24 22:55:02 +08:00
< img src = "https://img.shields.io/pypi/v/litellm.svg" alt = "PyPI Version" >
< / a >
2023-08-24 23:32:50 +08:00
< a href = "https://dl.circleci.com/status-badge/redirect/gh/BerriAI/litellm/tree/main" target = "_blank" >
2023-08-24 22:55:02 +08:00
< img src = "https://dl.circleci.com/status-badge/img/gh/BerriAI/litellm/tree/main.svg?style=svg" alt = "CircleCI" >
< / a >
2023-08-27 05:04:51 +08:00
< a href = "https://www.ycombinator.com/companies/berriai" >
< img src = "https://img.shields.io/badge/Y%20Combinator-W23-orange?style=flat-square" alt = "Y Combinator W23" >
< / a >
2023-10-18 04:57:41 +08:00
< a href = "https://wa.link/huol9n" >
< img src = "https://img.shields.io/static/v1?label=Chat%20on&message=WhatsApp&color=success&logo=WhatsApp&style=flat-square" alt = "Whatsapp" >
< / a >
< a href = "https://discord.gg/wuPM9dRgDw" >
< img src = "https://img.shields.io/static/v1?label=Chat%20on&message=Discord&color=blue&logo=Discord&style=flat-square" alt = "Discord" >
< / a >
2023-08-24 22:55:02 +08:00
< / h4 >
2023-09-27 23:03:39 +08:00
2023-12-26 14:57:37 +08:00
LiteLLM manages:
2024-02-28 14:07:13 +08:00
2024-01-09 15:17:08 +08:00
- Translate inputs to provider's `completion` , `embedding` , and `image_generation` endpoints
2023-12-26 14:57:37 +08:00
- [Consistent output ](https://docs.litellm.ai/docs/completion/output ), text responses will always be available at `['choices'][0]['message']['content']`
2024-01-09 15:18:49 +08:00
- Retry/fallback logic across multiple deployments (e.g. Azure/OpenAI) - [Router ](https://docs.litellm.ai/docs/routing )
2024-08-09 08:06:29 +08:00
- Set Budgets & Rate limits per project, api key, model [LiteLLM Proxy Server (LLM Gateway) ](https://docs.litellm.ai/docs/simple_proxy )
2024-02-10 01:25:36 +08:00
2024-08-09 08:06:29 +08:00
[**Jump to LiteLLM Proxy (LLM Gateway) Docs** ](https://github.com/BerriAI/litellm?tab=readme-ov-file#openai-proxy---docs ) < br >
2024-04-04 22:04:07 +08:00
[**Jump to Supported LLM Providers** ](https://github.com/BerriAI/litellm?tab=readme-ov-file#supported-providers-docs )
2023-09-03 08:42:59 +08:00
2024-05-19 10:06:09 +08:00
🚨 **Stable Release:** Use docker images with the `-stable` tag. These have undergone 12 hour load tests, before being published.
2024-03-27 01:34:16 +08:00
2024-02-14 09:19:41 +08:00
Support for more providers. Missing a provider or LLM Platform, raise a [feature request ](https://github.com/BerriAI/litellm/issues/new?assignees=&labels=enhancement&projects=&template=feature_request.yml&title=%5BFeature%5D%3A+ ).
2023-12-26 14:59:44 +08:00
# Usage ([**Docs**](https://docs.litellm.ai/docs/))
2024-02-28 14:07:13 +08:00
2023-11-12 07:41:20 +08:00
> [!IMPORTANT]
2024-06-26 00:24:00 +08:00
> LiteLLM v1.0.0 now requires `openai>=1.0.0`. Migration guide [here](https://docs.litellm.ai/docs/migration)
2024-06-26 00:23:19 +08:00
> LiteLLM v1.40.14+ now requires `pydantic>=2.0.0`. No changes required.
2023-11-12 07:41:20 +08:00
2023-10-15 01:32:29 +08:00
< a target = "_blank" href = "https://colab.research.google.com/github/BerriAI/litellm/blob/main/cookbook/liteLLM_Getting_Started.ipynb" >
2023-08-24 23:19:57 +08:00
< img src = "https://colab.research.google.com/assets/colab-badge.svg" alt = "Open In Colab" / >
< / a >
2023-09-27 23:04:37 +08:00
2023-12-27 20:00:37 +08:00
```shell
2023-07-31 22:58:10 +08:00
pip install litellm
```
2023-07-27 08:44:43 +08:00
```python
2023-07-27 08:46:23 +08:00
from litellm import completion
2023-08-30 01:39:33 +08:00
import os
2023-09-20 08:10:55 +08:00
2024-02-28 14:07:13 +08:00
## set ENV variables
os.environ["OPENAI_API_KEY"] = "your-openai-key"
os.environ["COHERE_API_KEY"] = "your-cohere-key"
2023-07-27 08:46:23 +08:00
2023-07-27 08:44:34 +08:00
messages = [{ "content": "Hello, how are you?","role": "user"}]
2023-07-27 08:46:23 +08:00
2023-07-27 08:44:34 +08:00
# openai call
response = completion(model="gpt-3.5-turbo", messages=messages)
# cohere call
2023-08-25 00:42:58 +08:00
response = completion(model="command-nightly", messages=messages)
2023-09-20 08:04:54 +08:00
print(response)
2023-07-27 08:44:34 +08:00
```
2023-08-01 23:18:44 +08:00
2024-02-27 07:23:52 +08:00
Call any model supported by a provider, with `model=<provider_name>/<model_name>` . There might be provider-specific details here, so refer to [provider docs for more information ](https://docs.litellm.ai/docs/providers )
2024-02-27 07:22:50 +08:00
2023-12-12 00:30:55 +08:00
## Async ([Docs](https://docs.litellm.ai/docs/completion/stream#async-completion))
```python
from litellm import acompletion
import asyncio
async def test_get_response():
user_message = "Hello, how are you?"
messages = [{"content": user_message, "role": "user"}]
response = await acompletion(model="gpt-3.5-turbo", messages=messages)
return response
response = asyncio.run(test_get_response())
print(response)
```
2023-10-01 00:48:45 +08:00
## Streaming ([Docs](https://docs.litellm.ai/docs/completion/stream))
2024-02-28 14:07:13 +08:00
2023-10-22 05:45:51 +08:00
liteLLM supports streaming the model response back, pass `stream=True` to get a streaming iterator in response.
2023-10-22 05:45:31 +08:00
Streaming is supported for all models (Bedrock, Huggingface, TogetherAI, Azure, OpenAI, etc.)
2024-02-28 14:07:13 +08:00
2023-08-06 06:12:34 +08:00
```python
2023-10-29 03:25:58 +08:00
from litellm import completion
2023-08-06 06:10:49 +08:00
response = completion(model="gpt-3.5-turbo", messages=messages, stream=True)
2023-11-22 03:19:32 +08:00
for part in response:
print(part.choices[0].delta.content or "")
2023-08-09 07:22:18 +08:00
2023-08-09 07:19:34 +08:00
# claude 2
2023-11-22 03:19:32 +08:00
response = completion('claude-2', messages, stream=True)
for part in response:
print(part.choices[0].delta.content or "")
2023-08-06 06:10:49 +08:00
```
2023-10-29 03:29:44 +08:00
2023-11-09 05:37:02 +08:00
## Logging Observability ([Docs](https://docs.litellm.ai/docs/observability/callbacks))
2024-02-28 14:07:13 +08:00
LiteLLM exposes pre defined callbacks to send data to Lunary, Langfuse, DynamoDB, s3 Buckets, Helicone, Promptlayer, Traceloop, Athina, Slack
2023-10-29 03:29:44 +08:00
```python
from litellm import completion
## set env variables for logging tools
2024-03-19 11:50:04 +08:00
os.environ["LUNARY_PUBLIC_KEY"] = "your-lunary-public-key"
2024-07-09 06:41:35 +08:00
os.environ["HELICONE_API_KEY"] = "your-helicone-auth-key"
2023-11-19 06:52:24 +08:00
os.environ["LANGFUSE_PUBLIC_KEY"] = ""
os.environ["LANGFUSE_SECRET_KEY"] = ""
2024-02-24 00:19:40 +08:00
os.environ["ATHINA_API_KEY"] = "your-athina-api-key"
2023-10-29 03:29:44 +08:00
os.environ["OPENAI_API_KEY"]
# set callbacks
2024-07-09 06:41:35 +08:00
litellm.success_callback = ["lunary", "langfuse", "athina", "helicone"] # log input/output to lunary, langfuse, supabase, athina, helicone etc
2023-10-29 03:29:44 +08:00
#openai call
response = completion(model="gpt-3.5-turbo", messages=[{"role": "user", "content": "Hi 👋 - i'm openai"}])
2023-10-29 03:25:58 +08:00
```
2024-08-09 08:06:29 +08:00
# LiteLLM Proxy Server (LLM Gateway) - ([Docs](https://docs.litellm.ai/docs/simple_proxy))
2023-12-26 14:57:37 +08:00
2024-04-21 00:48:49 +08:00
Track spend + Load Balance across multiple projects
[Hosted Proxy (Preview) ](https://docs.litellm.ai/docs/hosted )
2023-12-26 14:57:37 +08:00
2024-02-28 14:07:13 +08:00
The proxy provides:
2024-01-02 19:52:44 +08:00
1. [Hooks for auth ](https://docs.litellm.ai/docs/proxy/virtual_keys#custom-auth )
2. [Hooks for logging ](https://docs.litellm.ai/docs/proxy/logging#step-1---create-your-custom-litellm-callback-class )
3. [Cost tracking ](https://docs.litellm.ai/docs/proxy/virtual_keys#tracking-spend )
4. [Rate Limiting ](https://docs.litellm.ai/docs/proxy/users#set-rate-limits )
2023-12-28 14:28:33 +08:00
## 📖 Proxy Endpoints - [Swagger Docs](https://litellm-api.up.railway.app/)
2024-06-02 07:53:06 +08:00
2024-02-28 14:07:13 +08:00
## Quick Start Proxy - CLI
2023-12-27 20:00:37 +08:00
```shell
2024-01-06 16:19:15 +08:00
pip install 'litellm[proxy]'
2023-12-27 20:00:37 +08:00
```
2023-12-26 14:57:37 +08:00
### Step 1: Start litellm proxy
2024-02-28 14:07:13 +08:00
2023-12-26 14:57:37 +08:00
```shell
$ litellm --model huggingface/bigcode/starcoder
2024-03-09 13:59:00 +08:00
#INFO: Proxy running on http://0.0.0.0:4000
2023-12-26 14:57:37 +08:00
```
2023-12-27 14:23:35 +08:00
### Step 2: Make ChatCompletions Request to Proxy
2024-02-28 14:07:13 +08:00
2024-07-26 11:10:02 +08:00
> [!IMPORTANT]
2024-07-26 11:12:32 +08:00
> 💡 [Use LiteLLM Proxy with Langchain (Python, JS), OpenAI SDK (Python, JS) Anthropic SDK, Mistral SDK, LlamaIndex, Instructor, Curl](https://docs.litellm.ai/docs/proxy/user_keys)
2024-07-26 11:10:02 +08:00
2023-12-26 14:57:37 +08:00
```python
import openai # openai v1.0.0+
2024-03-09 13:59:00 +08:00
client = openai.OpenAI(api_key="anything",base_url="http://0.0.0.0:4000") # set proxy to base_url
2023-12-26 14:57:37 +08:00
# 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)
```
2023-12-27 14:23:35 +08:00
## Proxy Key Management ([Docs](https://docs.litellm.ai/docs/proxy/virtual_keys))
2024-02-28 14:07:13 +08:00
2024-06-02 07:53:06 +08:00
Connect the proxy with a Postgres DB to create proxy keys
```bash
# Get the code
git clone https://github.com/BerriAI/litellm
# Go to folder
cd litellm
2024-07-19 04:08:06 +08:00
# Add the master key - you can change this after setup
2024-06-02 07:53:06 +08:00
echo 'LITELLM_MASTER_KEY="sk-1234"' > .env
2024-07-19 04:08:06 +08:00
# Add the litellm salt key - you cannot change this after adding a model
# It is used to encrypt / decrypt your LLM API Key credentials
# We recommned - https://1password.com/password-generator/
# password generator to get a random hash for litellm salt key
echo 'LITELLM_SALT_KEY="sk-1234"' > .env
2024-06-02 07:53:06 +08:00
source .env
# Start
docker-compose up
```
2024-02-28 14:07:13 +08:00
UI on `/ui` on your proxy server
2024-02-10 01:30:16 +08:00
2024-02-25 02:04:47 +08:00
Set budgets and rate limits across multiple projects
2023-12-27 14:23:35 +08:00
`POST /key/generate`
2023-12-27 14:31:05 +08:00
### Request
2024-02-28 14:07:13 +08:00
2023-12-27 14:23:35 +08:00
```shell
2024-03-09 13:59:00 +08:00
curl 'http://0.0.0.0:4000/key/generate' \
2023-12-27 14:23:35 +08:00
--header 'Authorization: Bearer sk-1234' \
--header 'Content-Type: application/json' \
--data-raw '{"models": ["gpt-3.5-turbo", "gpt-4", "claude-2"], "duration": "20m","metadata": {"user": "ishaan@berri.ai", "team": "core-infra"}}'
```
2023-12-27 14:37:25 +08:00
### Expected Response
2024-02-28 14:07:13 +08:00
2023-12-27 14:31:05 +08:00
```shell
{
"key": "sk-kdEXbIqZRwEeEiHwdg7sFA", # Bearer token
"expires": "2023-11-19T01:38:25.838000+00:00" # datetime object
}
```
2024-01-03 17:47:24 +08:00
## Supported Providers ([Docs](https://docs.litellm.ai/docs/providers))
2024-01-09 15:21:30 +08:00
2024-02-28 14:07:13 +08:00
| Provider | [Completion ](https://docs.litellm.ai/docs/#basic-usage ) | [Streaming ](https://docs.litellm.ai/docs/completion/stream#streaming-responses ) | [Async Completion ](https://docs.litellm.ai/docs/completion/stream#async-completion ) | [Async Streaming ](https://docs.litellm.ai/docs/completion/stream#async-streaming ) | [Async Embedding ](https://docs.litellm.ai/docs/embedding/supported_embedding ) | [Async Image Generation ](https://docs.litellm.ai/docs/image_generation ) |
2024-06-06 20:56:49 +08:00
|-------------------------------------------------------------------------------------|---------------------------------------------------------|---------------------------------------------------------------------------------|-------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------|-------------------------------------------------------------------------------|-------------------------------------------------------------------------|
| [openai ](https://docs.litellm.ai/docs/providers/openai ) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| [azure ](https://docs.litellm.ai/docs/providers/azure ) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| [aws - sagemaker ](https://docs.litellm.ai/docs/providers/aws_sagemaker ) | ✅ | ✅ | ✅ | ✅ | ✅ | |
| [aws - bedrock ](https://docs.litellm.ai/docs/providers/bedrock ) | ✅ | ✅ | ✅ | ✅ | ✅ | |
| [google - vertex_ai ](https://docs.litellm.ai/docs/providers/vertex ) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| [google - palm ](https://docs.litellm.ai/docs/providers/palm ) | ✅ | ✅ | ✅ | ✅ | | |
| [google AI Studio - gemini ](https://docs.litellm.ai/docs/providers/gemini ) | ✅ | ✅ | ✅ | ✅ | | |
| [mistral ai api ](https://docs.litellm.ai/docs/providers/mistral ) | ✅ | ✅ | ✅ | ✅ | ✅ | |
| [cloudflare AI Workers ](https://docs.litellm.ai/docs/providers/cloudflare_workers ) | ✅ | ✅ | ✅ | ✅ | | |
| [cohere ](https://docs.litellm.ai/docs/providers/cohere ) | ✅ | ✅ | ✅ | ✅ | ✅ | |
| [anthropic ](https://docs.litellm.ai/docs/providers/anthropic ) | ✅ | ✅ | ✅ | ✅ | | |
2024-06-01 09:55:22 +08:00
| [empower ](https://docs.litellm.ai/docs/providers/empower ) | ✅ | ✅ | ✅ | ✅ |
2024-06-06 20:56:49 +08:00
| [huggingface ](https://docs.litellm.ai/docs/providers/huggingface ) | ✅ | ✅ | ✅ | ✅ | ✅ | |
| [replicate ](https://docs.litellm.ai/docs/providers/replicate ) | ✅ | ✅ | ✅ | ✅ | | |
| [together_ai ](https://docs.litellm.ai/docs/providers/togetherai ) | ✅ | ✅ | ✅ | ✅ | | |
| [openrouter ](https://docs.litellm.ai/docs/providers/openrouter ) | ✅ | ✅ | ✅ | ✅ | | |
| [ai21 ](https://docs.litellm.ai/docs/providers/ai21 ) | ✅ | ✅ | ✅ | ✅ | | |
| [baseten ](https://docs.litellm.ai/docs/providers/baseten ) | ✅ | ✅ | ✅ | ✅ | | |
| [vllm ](https://docs.litellm.ai/docs/providers/vllm ) | ✅ | ✅ | ✅ | ✅ | | |
| [nlp_cloud ](https://docs.litellm.ai/docs/providers/nlp_cloud ) | ✅ | ✅ | ✅ | ✅ | | |
| [aleph alpha ](https://docs.litellm.ai/docs/providers/aleph_alpha ) | ✅ | ✅ | ✅ | ✅ | | |
| [petals ](https://docs.litellm.ai/docs/providers/petals ) | ✅ | ✅ | ✅ | ✅ | | |
| [ollama ](https://docs.litellm.ai/docs/providers/ollama ) | ✅ | ✅ | ✅ | ✅ | ✅ | |
| [deepinfra ](https://docs.litellm.ai/docs/providers/deepinfra ) | ✅ | ✅ | ✅ | ✅ | | |
| [perplexity-ai ](https://docs.litellm.ai/docs/providers/perplexity ) | ✅ | ✅ | ✅ | ✅ | | |
| [Groq AI ](https://docs.litellm.ai/docs/providers/groq ) | ✅ | ✅ | ✅ | ✅ | | |
| [Deepseek ](https://docs.litellm.ai/docs/providers/deepseek ) | ✅ | ✅ | ✅ | ✅ | | |
| [anyscale ](https://docs.litellm.ai/docs/providers/anyscale ) | ✅ | ✅ | ✅ | ✅ | | |
| [IBM - watsonx.ai ](https://docs.litellm.ai/docs/providers/watsonx ) | ✅ | ✅ | ✅ | ✅ | ✅ | |
| [voyage ai ](https://docs.litellm.ai/docs/providers/voyage ) | | | | | ✅ | |
| [xinference [Xorbits Inference]](https://docs.litellm.ai/docs/providers/xinference) | | | | | ✅ | |
2024-06-11 15:23:08 +08:00
| [FriendliAI ](https://docs.litellm.ai/docs/providers/friendliai ) | ✅ | ✅ | ✅ | ✅ | | |
2023-10-01 00:48:45 +08:00
[**Read the Docs** ](https://docs.litellm.ai/docs/ )
2023-10-18 04:38:55 +08:00
## Contributing
2023-09-19 00:30:27 +08:00
2024-02-28 14:07:13 +08:00
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
2023-09-19 00:30:27 +08:00
```
git clone https://github.com/BerriAI/litellm.git
```
2024-02-28 14:07:13 +08:00
Step 2: Navigate into the project, and install dependencies:
2023-09-19 00:30:27 +08:00
```
cd litellm
2024-04-23 02:35:08 +08:00
poetry install -E extra_proxy -E proxy
2023-09-19 00:30:27 +08:00
```
Step 3: Test your change:
2024-02-28 14:07:13 +08:00
2023-09-19 00:30:27 +08:00
```
cd litellm/tests # pwd: Documents/litellm/litellm/tests
2023-12-22 16:28:37 +08:00
poetry run flake8
2023-12-22 16:25:57 +08:00
poetry run pytest .
2023-09-19 00:30:27 +08:00
```
Step 4: Submit a PR with your changes! 🚀
2024-02-28 14:07:13 +08:00
- push your fork to your GitHub repo
- submit a PR from there
2023-09-19 00:30:27 +08:00
2024-03-03 09:18:48 +08:00
# Enterprise
For companies that need better security, user management and professional support
[Talk to founders ](https://calendly.com/d/4mp-gd3-k5k/litellm-1-1-onboarding-chat )
This covers:
- ✅ **Features under the [LiteLLM Commercial License](https://docs.litellm.ai/docs/proxy/enterprise):**
- ✅ **Feature Prioritization**
- ✅ **Custom Integrations**
- ✅ **Professional Support - Dedicated discord + slack**
- ✅ **Custom SLAs**
- ✅ **Secure access with Single Sign-On**
2023-08-28 23:59:34 +08:00
# Support / talk with founders
2024-02-28 14:07:13 +08:00
2023-08-24 23:43:18 +08:00
- [Schedule Demo 👋 ](https://calendly.com/d/4mp-gd3-k5k/berriai-1-1-onboarding-litellm-hosted-version )
2023-08-09 21:09:28 +08:00
- [Community Discord 💭 ](https://discord.gg/wuPM9dRgDw )
- Our numbers 📞 +1 (770) 8783-106 /
- Our emails ✉️ ishaan@berri.ai / krrish@berri.ai
2023-07-29 00:11:27 +08:00
2024-02-28 14:07:13 +08:00
# Why did we build this
2023-10-17 02:40:46 +08:00
- **Need for simplicity**: Our code started to get extremely complicated managing & translating calls between Azure, OpenAI and Cohere.
2023-08-28 23:50:12 +08:00
2023-08-28 23:59:34 +08:00
# Contributors
2023-08-28 23:50:12 +08:00
<!-- ALL - CONTRIBUTORS - LIST:START - Do not remove or modify this section -->
<!-- prettier - ignore - start -->
<!-- markdownlint - disable -->
<!-- markdownlint - restore -->
<!-- prettier - ignore - end -->
<!-- ALL - CONTRIBUTORS - LIST:END -->
2023-08-28 23:56:19 +08:00
< a href = "https://github.com/BerriAI/litellm/graphs/contributors" >
< img src = "https://contrib.rocks/image?repo=BerriAI/litellm" / >
< / a >
