Skip to main content

Caching

Cache LLM Responses

note

For OpenAI/Anthropic Prompt Caching, go here

LiteLLM supports:

  • In Memory Cache
  • Redis Cache
  • Qdrant Semantic Cache
  • Redis Semantic Cache
  • s3 Bucket Cache

Quick Start - Redis, s3 Cache, Semantic Cache​

Caching can be enabled by adding the cache key in the config.yaml

Step 1: Add cache to the config.yaml​

model_list:
  - model_name: gpt-3.5-turbo
    litellm_params:
      model: gpt-3.5-turbo
  - model_name: text-embedding-ada-002
    litellm_params:
      model: text-embedding-ada-002

litellm_settings:
  set_verbose: True
  cache: True          # set cache responses to True, litellm defaults to using a redis cache

[OPTIONAL] Step 1.5: Add redis namespaces, default ttl​

Namespace​

If you want to create some folder for your keys, you can set a namespace, like this:

litellm_settings:
  cache: true 
  cache_params:        # set cache params for redis
    type: redis
    namespace: "litellm.caching.caching"

and keys will be stored like:

litellm.caching.caching:<hash>

Redis Cluster​

model_list:
  - model_name: "*"
    litellm_params:
      model: "*"


litellm_settings:
  cache: True
  cache_params:
    type: redis
    redis_startup_nodes: [{"host": "127.0.0.1", "port": "7001"}]

Redis Sentinel​

model_list:
  - model_name: "*"
    litellm_params:
      model: "*"


litellm_settings:
  cache: true
  cache_params:
    type: "redis"
    service_name: "mymaster"
    sentinel_nodes: [["localhost", 26379]]

TTL​

litellm_settings:
  cache: true 
  cache_params:        # set cache params for redis
    type: redis
    ttl: 600 # will be cached on redis for 600s
    # default_in_memory_ttl: Optional[float], default is None. time in seconds. 
    # default_in_redis_ttl: Optional[float], default is None. time in seconds.

SSL​

just set REDIS_SSL="True" in your .env, and LiteLLM will pick this up. 

REDIS_SSL="True"

For quick testing, you can also use REDIS_URL, eg.:

REDIS_URL="rediss://.."

but we don't recommend using REDIS_URL in prod. We've noticed a performance difference between using it vs. redis_host, port, etc.

Step 2: Add Redis Credentials to .env​

Set either REDIS_URL or the REDIS_HOST in your os environment, to enable caching.

REDIS_URL = ""        # REDIS_URL='redis://username:password@hostname:port/database'
## OR ## 
REDIS_HOST = ""       # REDIS_HOST='redis-18841.c274.us-east-1-3.ec2.cloud.redislabs.com'
REDIS_PORT = ""       # REDIS_PORT='18841'
REDIS_PASSWORD = ""   # REDIS_PASSWORD='liteLlmIsAmazing'

Additional kwargs
You can pass in any additional redis.Redis arg, by storing the variable + value in your os environment, like this: 

REDIS_<redis-kwarg-name> = ""

See how it's read from the environment

Step 3: Run proxy with config​

$ litellm --config /path/to/config.yaml

Using Caching - /chat/completions​

Send the same request twice:

curl http://0.0.0.0:4000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
     "model": "gpt-3.5-turbo",
     "messages": [{"role": "user", "content": "write a poem about litellm!"}],
     "temperature": 0.7
   }'

curl http://0.0.0.0:4000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
     "model": "gpt-3.5-turbo",
     "messages": [{"role": "user", "content": "write a poem about litellm!"}],
     "temperature": 0.7
   }'

Set cache for proxy, but not on the actual llm api call​

Use this if you just want to enable features like rate limiting, and loadbalancing across multiple instances.

Set supported_call_types: [] to disable caching on the actual api call. 

litellm_settings:
  cache: True
  cache_params:
    type: redis
    supported_call_types: []

Debugging Caching - /cache/ping​

LiteLLM Proxy exposes a /cache/ping endpoint to test if the cache is working as expected

Usage

curl --location 'http://0.0.0.0:4000/cache/ping'  -H "Authorization: Bearer sk-1234"

Expected Response - when cache healthy

{
    "status": "healthy",
    "cache_type": "redis",
    "ping_response": true,
    "set_cache_response": "success",
    "litellm_cache_params": {
        "supported_call_types": "['completion', 'acompletion', 'embedding', 'aembedding', 'atranscription', 'transcription']",
        "type": "redis",
        "namespace": "None"
    },
    "redis_cache_params": {
        "redis_client": "Redis<ConnectionPool<Connection<host=redis-16337.c322.us-east-1-2.ec2.cloud.redislabs.com,port=16337,db=0>>>",
        "redis_kwargs": "{'url': 'redis://:******@redis-16337.c322.us-east-1-2.ec2.cloud.redislabs.com:16337'}",
        "async_redis_conn_pool": "BlockingConnectionPool<Connection<host=redis-16337.c322.us-east-1-2.ec2.cloud.redislabs.com,port=16337,db=0>>",
        "redis_version": "7.2.0"
    }
}

Advanced​

Control Call Types Caching is on for - (/chat/completion, /embeddings, etc.)​

By default, caching is on for all call types. You can control which call types caching is on for by setting supported_call_types in cache_params

Cache will only be on for the call types specified in supported_call_types

litellm_settings:
  cache: True
  cache_params:
    type: redis
    supported_call_types: ["acompletion", "atext_completion", "aembedding", "atranscription"]
                          # /chat/completions, /completions, /embeddings, /audio/transcriptions

Set Cache Params on config.yaml​

model_list:
  - model_name: gpt-3.5-turbo
    litellm_params:
      model: gpt-3.5-turbo
  - model_name: text-embedding-ada-002
    litellm_params:
      model: text-embedding-ada-002

litellm_settings:
  set_verbose: True
  cache: True          # set cache responses to True, litellm defaults to using a redis cache
  cache_params:         # cache_params are optional
    type: "redis"  # The type of cache to initialize. Can be "local" or "redis". Defaults to "local".
    host: "localhost"  # The host address for the Redis cache. Required if type is "redis".
    port: 6379  # The port number for the Redis cache. Required if type is "redis".
    password: "your_password"  # The password for the Redis cache. Required if type is "redis".
    
    # Optional configurations
    supported_call_types: ["acompletion", "atext_completion", "aembedding", "atranscription"]
                      # /chat/completions, /completions, /embeddings, /audio/transcriptions

Turn on / off caching per request. ​

The proxy support 4 cache-controls:

  • ttl: Optional(int) - Will cache the response for the user-defined amount of time (in seconds).
  • s-maxage: Optional(int) Will only accept cached responses that are within user-defined range (in seconds).
  • no-cache: Optional(bool) Will not return a cached response, but instead call the actual endpoint.
  • no-store: Optional(bool) Will not cache the response.

Let us know if you need more

Turn off caching

Set no-cache=True, this will not return a cached response

import os
from openai import OpenAI

client = OpenAI(
    # This is the default and can be omitted
    api_key=os.environ.get("OPENAI_API_KEY"),
        base_url="http://0.0.0.0:4000"
)

chat_completion = client.chat.completions.create(
    messages=[
        {
            "role": "user",
            "content": "Say this is a test",
        }
    ],
    model="gpt-3.5-turbo",
    extra_body = {        # OpenAI python accepts extra args in extra_body
        cache: {
          "no-cache": True # will not return a cached response 
      }
    }
)

Turn on caching

By default cache is always on

import os
from openai import OpenAI

client = OpenAI(
    # This is the default and can be omitted
    api_key=os.environ.get("OPENAI_API_KEY"),
        base_url="http://0.0.0.0:4000"
)

chat_completion = client.chat.completions.create(
    messages=[
        {
            "role": "user",
            "content": "Say this is a test",
        }
    ],
    model="gpt-3.5-turbo"
)

Set ttl

Set ttl=600, this will caches response for 10 minutes (600 seconds)

import os
from openai import OpenAI

client = OpenAI(
    # This is the default and can be omitted
    api_key=os.environ.get("OPENAI_API_KEY"),
        base_url="http://0.0.0.0:4000"
)

chat_completion = client.chat.completions.create(
    messages=[
        {
            "role": "user",
            "content": "Say this is a test",
        }
    ],
    model="gpt-3.5-turbo",
    extra_body = {        # OpenAI python accepts extra args in extra_body
        cache: {
          "ttl": 600 # caches response for 10 minutes 
      }
    }
)

Set s-maxage

Set s-maxage, this will only get responses cached within last 10 minutes 

import os
from openai import OpenAI

client = OpenAI(
    # This is the default and can be omitted
    api_key=os.environ.get("OPENAI_API_KEY"),
        base_url="http://0.0.0.0:4000"
)

chat_completion = client.chat.completions.create(
    messages=[
        {
            "role": "user",
            "content": "Say this is a test",
        }
    ],
    model="gpt-3.5-turbo",
    extra_body = {        # OpenAI python accepts extra args in extra_body
        cache: {
          "s-maxage": 600 # only get responses cached within last 10 minutes 
      }
    }
)

Turn on / off caching per Key.​

  1. Add cache params when creating a key full list
curl -X POST 'http://0.0.0.0:4000/key/generate' \
-H 'Authorization: Bearer sk-1234' \
-H 'Content-Type: application/json' \
-d '{
    "user_id": "222",
    "metadata": {
        "cache": {
            "no-cache": true
        }
    }
}'
  1. Test it! 
curl -X POST 'http://localhost:4000/chat/completions' \
-H 'Content-Type: application/json' \
-H 'Authorization: Bearer <YOUR_NEW_KEY>' \
-d '{"model": "gpt-3.5-turbo", "messages": [{"role": "user", "content": "bom dia"}]}'

Deleting Cache Keys - /cache/delete​

In order to delete a cache key, send a request to /cache/delete with the keys you want to delete

Example 

curl -X POST "http://0.0.0.0:4000/cache/delete" \
  -H "Authorization: Bearer sk-1234" \
  -d '{"keys": ["586bf3f3c1bf5aecb55bd9996494d3bbc69eb58397163add6d49537762a7548d", "key2"]}'
# {"status":"success"}

Viewing Cache Keys from responses​

You can view the cache_key in the response headers, on cache hits the cache key is sent as the x-litellm-cache-key response headers

curl -i --location 'http://0.0.0.0:4000/chat/completions' \
    --header 'Authorization: Bearer sk-1234' \
    --header 'Content-Type: application/json' \
    --data '{
    "model": "gpt-3.5-turbo",
    "user": "ishan",
    "messages": [
        {
        "role": "user",
        "content": "what is litellm"
        }
    ],
}'

Response from litellm proxy 

date: Thu, 04 Apr 2024 17:37:21 GMT
content-type: application/json
x-litellm-cache-key: 586bf3f3c1bf5aecb55bd9996494d3bbc69eb58397163add6d49537762a7548d

{
    "id": "chatcmpl-9ALJTzsBlXR9zTxPvzfFFtFbFtG6T",
    "choices": [
        {
            "finish_reason": "stop",
            "index": 0,
            "message": {
                "content": "I'm sorr.."
                "role": "assistant"
            }
        }
    ],
    "created": 1712252235,
}

Set Caching Default Off - Opt in only ​

  1. Set mode: default_off for caching
model_list:
  - model_name: fake-openai-endpoint
    litellm_params:
      model: openai/fake
      api_key: fake-key
      api_base: https://exampleopenaiendpoint-production.up.railway.app/

# default off mode
litellm_settings:
  set_verbose: True
  cache: True
  cache_params:
    mode: default_off # 👈 Key change cache is default_off
  1. Opting in to cache when cache is default off
import os
from openai import OpenAI

client = OpenAI(api_key=<litellm-api-key>, base_url="http://0.0.0.0:4000")

chat_completion = client.chat.completions.create(
    messages=[
        {
            "role": "user",
            "content": "Say this is a test",
        }
    ],
    model="gpt-3.5-turbo",
    extra_body = {        # OpenAI python accepts extra args in extra_body
        "cache": {"use-cache": True}
    }
)

Turn on batch_redis_requests​

What it does? When a request is made:

  • Check if a key starting with litellm:<hashed_api_key>:<call_type>: exists in-memory, if no - get the last 100 cached requests for this key and store it

  • New requests are stored with this litellm:.. as the namespace

Why? Reduce number of redis GET requests. This improved latency by 46% in prod load tests.

Usage

litellm_settings:
  cache: true
  cache_params:
    type: redis
    ... # remaining redis args (host, port, etc.)
  callbacks: ["batch_redis_requests"] # 👈 KEY CHANGE!

SEE CODE

Supported cache_params on proxy config.yaml​

cache_params:
  # ttl 
  ttl: Optional[float]
  default_in_memory_ttl: Optional[float]
  default_in_redis_ttl: Optional[float]

  # Type of cache (options: "local", "redis", "s3")
  type: s3

  # List of litellm call types to cache for
  # Options: "completion", "acompletion", "embedding", "aembedding"
  supported_call_types: ["acompletion", "atext_completion", "aembedding", "atranscription"]
                      # /chat/completions, /completions, /embeddings, /audio/transcriptions

  # Redis cache parameters
  host: localhost  # Redis server hostname or IP address
  port: "6379"  # Redis server port (as a string)
  password: secret_password  # Redis server password
  namespace: Optional[str] = None,
  

  # S3 cache parameters
  s3_bucket_name: your_s3_bucket_name  # Name of the S3 bucket
  s3_region_name: us-west-2  # AWS region of the S3 bucket
  s3_api_version: 2006-03-01  # AWS S3 API version
  s3_use_ssl: true  # Use SSL for S3 connections (options: true, false)
  s3_verify: true  # SSL certificate verification for S3 connections (options: true, false)
  s3_endpoint_url: https://s3.amazonaws.com  # S3 endpoint URL
  s3_aws_access_key_id: your_access_key  # AWS Access Key ID for S3
  s3_aws_secret_access_key: your_secret_key  # AWS Secret Access Key for S3
  s3_aws_session_token: your_session_token  # AWS Session Token for temporary credentials

Advanced - user api key cache ttl​

Configure how long the in-memory cache stores the key object (prevents db requests)

general_settings:
  user_api_key_cache_ttl: <your-number> #time in seconds

By default this value is set to 60s.