R package providing a simple interface to interact with LiteLLM AI language models.

✈️ Installation

Run these R commands:

# Install remotes package if necessary
install.packages("remotes")
# Install litellmR package from GitHub
remotes::install_github("vdwulp/litellmR")

🚶 Basic usage

Code example:

library(litellmR)
# Set up the connection
litellm_setup(
  api_key = "YOUR_API_KEY",
  base_url = "https://my-litellm-server/v1"
)
# 2. Check available models
litellm_models()
# 3. Send a prompt
litellm_prompt("Explain regression analysis in simple terms.")

For extended documentation, use:

?litellmR

🏃 Advanced usage

Multiple prompts at once

Prompts can be provided as a vector:

prompts <- c(
  "Summarize key points about teaching feedback.",
  "Give tips for improving student engagement."
)
litellm_prompt(prompts)

Or integrated in a tidyverse pipeline:

df <- tibble(
  Number = 1:2,
  Request = c( "Summarize key points about teaching feedback.",
               "Give tips for improving student engagement." )
)
df |>
  mutate(Response = litellm_prompt(Request))

Each element in the vector or data frame column is treated as a separate prompt.

Multi-turn chat

For a multi-turn chat where context is preserved across messages:

litellm_chat("Explain regression analysis in simple terms.")
litellm_chat("That is way too complex for me.")

litellm_prompt() is stateless, best for single or batched prompts where context is not needed.
litellm_chat() preserves conversation context, suited for multi-turn interactions.

Model and temperature

You can specify the AI model and/or temperature:

litellm_prompt("Explain regression analysis in simple terms.",
               model = "gpt-4.1",
               temperature = 1.0 )

model selects which LiteLLM AI model to use (default gpt-4o-mini, gpt-4.1 in example).
temperature controls creativity/randomness (0 = deterministic, 1 = creative, default 0.7).

🗒️ License

MIT License, full text available in LICENSE file.