Skip to main content

LiteLLM Inside

HeaderFor
Request
x-litellm-timeout: <seconds>请求超时设置
x-litellm-stream-timeout: <seconds>第一个 Chunk 超时
x-litellm-enable-message-redaction: <boolean>启用消息内容屏蔽
x-litellm-tags: <tag1,tag2,...>请求标签
x-litellm-num-retries: <number>请求重试次数
x-litellm-spend-logs-metadata: <json>请求开销日志元数据
Request/Anthropic
anthropic-version: <str>API version
anthropic-beta: <str>beta version
Request/OpenAI
openai-organization: <str>organization id
Request/Bypass
x-*需要配置 forward_client_headers_to_llm_api
Response/Rate Limit
x-ratelimit-remaining-requests: <int>剩余可用请求数
x-ratelimit-remaining-tokens: <int>剩余可用token数
x-ratelimit-limit-requests: <int>最大请求数限制
x-ratelimit-limit-tokens: <int>最大token数限制
x-ratelimit-reset-requests: <int>请求限制重置时间
x-ratelimit-reset-tokens: <int>token限制重置时间
Response/Latency
x-litellm-response-duration-ms: <float>从请求到响应的总耗时(毫秒)
x-litellm-overhead-duration-ms: <float>LiteLLM处理开销时间(毫秒)
Response/Retry&Fallback
x-litellm-attempted-retries: <int>实际重试次数
x-litellm-attempted-fallbacks: <int>实际回退次数
x-litellm-max-fallbacks: <int>最大回退次数限制
Response/Cost
x-litellm-response-cost: <float>API调用费用
x-litellm-key-spend: <float>API密钥总消费
Response/Bypass
llm_provider-*透传LLM提供商的响应头
spend-logs-metadata
{ "user_id": "12345", "project_id": "proj_abc", "request_type": "chat_completion" }

config.yaml​

include:
  - model_config.yaml

model_list: []
litellm_settings:
  num_retries: 3 # retry call 3 times on each model_name (e.g. zephyr-beta)
  request_timeout: 10 # raise Timeout error if call takes longer than 10s. Sets litellm.request_timeout
  fallbacks: [{"zephyr-beta": ["gpt-4o"]}] # fallback to gpt-4o if call fails num_retries
  context_window_fallbacks: [{"zephyr-beta": ["gpt-3.5-turbo-16k"]}, {"gpt-4o": ["gpt-3.5-turbo-16k"]}] # fallback to gpt-3.5-turbo-16k if context window error
  allowed_fails: 3 # cooldown model if it fails > 1 call in a minute.
router_settings: # router_settings are optional
  routing_strategy: simple-shuffle # Literal["simple-shuffle", "least-busy", "usage-based-routing","latency-based-routing"], default="simple-shuffle"
  model_group_alias: {"gpt-4": "gpt-4o"} # all requests with `gpt-4` will be routed to models with `gpt-4o`
  num_retries: 2
  timeout: 30                                  # 30 seconds
  redis_host: <your redis host>                # set this when using multiple litellm proxy deployments, load balancing state stored in redis
  redis_password: <your redis password>
  redis_port: 1992
general_settings: {}
environment_variables: {}
model_list:
  - model_name: glm-4.5
    litellm_params:
      model: openai/glm-4.5
      litellm_credential_name: zhipu_credential

  - model_name: glm-4.5-air
    litellm_params:
      model: openai/glm-4.5-air
      litellm_credential_name: zhipu_credential

  - model_name: '*'
    litellm_params:
      model: openai/glm-4.5-air
      litellm_credential_name: zhipu_credential

credential_list:
  - credential_name: zhipu_credential
    credential_values:
      api_base: os.environ/ZHIPU_API_BASE
      api_key: os.environ/ZHIPU_API_KEY
    credential_info:
      description: '智普'

支持通配符

model_list:
  - model_name: xai/*
    litellm_params:
      model: xai/*
      api_key: os.environ/XAI_API_KEY

litellm_settings:
  check_provider_endpoint: true
# params for litellm.completion() - https://docs.litellm.ai/docs/completion/input#input---request-body
litellm_params:
  model: openai/facebook/opt-125m
  api_base: http://0.0.0.0:4000/v1
  api_key: none
  api_version: "2023-05-15"
  rpm: 60      # Optional[int]: When rpm/tpm set - litellm uses weighted pick for load balancing. rpm = Rate limit for this deployment: in requests per minute (rpm).
  tpm: 1000   # Optional[int]: tpm = Tokens Per Minute
  azure_ad_token: ""
  seed: 1234
  max_token: 1024
  temperature: 0.2
  organization: "org-12345"
  aws_region_name: "us-west-2"
  extra_headers: {"AI-Resource Group": "ishaan-resource"}
model_info:
  version: 2
  access_groups: ['restricted-models']
  supported_environments: ["development", "production", "staging"]
  custom_tokenizer:
    identifier: deepseek-ai/DeepSeek-V3-Base
    revision: main
    auth_token: os.environ/HUGGINGFACE_API_KEY

参考​