LiteLLM

Provider value: openai-completions

LiteLLM is a proxy that provides a unified OpenAI-compatible API in front of 100+ LLM providers. Run the LiteLLM proxy locally or on your infrastructure, then point ra at it using the openai-completions provider.

Quick start

1. Start the LiteLLM proxy (see LiteLLM docs):

litellm --model claude-sonnet-4

2. Connect ra:

ra --provider openai-completions \
  --openai-base-url http://localhost:4000/v1 \
  --model claude-sonnet-4 "Hello"

Config file

app:
  providers:
    openai-completions:
      baseURL: http://localhost:4000/v1
      apiKey: ${LITELLM_API_KEY}

agent:
  provider: openai-completions
  model: claude-sonnet-4

With an API key

If your LiteLLM proxy requires authentication:

export OPENAI_API_KEY=sk-your-litellm-key
ra --provider openai-completions \
  --openai-base-url http://localhost:4000/v1 \
  --model claude-sonnet-4 "Hello"

Environment variables

Variable	Required	Description
OPENAI_API_KEY	Depends	Required if LiteLLM proxy has auth enabled

TIP

LiteLLM passes model-specific credentials (Anthropic, OpenAI, etc.) through its own configuration — ra only needs to authenticate with the proxy itself.

Remote proxy

Point ra at a LiteLLM proxy running on another machine:

app:
  providers:
    openai-completions:
      baseURL: https://litellm.your-company.com/v1
      apiKey: ${LITELLM_API_KEY}

agent:
  provider: openai-completions
  model: claude-sonnet-4

See also

• OpenAI Completions — the underlying provider
• Configuration — provider credentials reference