Simple chat with LLMR

knitr::opts_chunk$set(
  collapse = TRUE, comment = "#>",
  eval = identical(tolower(Sys.getenv("LLMR_RUN_VIGNETTES", "false")), "true")
)

This vignette shows basic chat usage with four providers and model names: - OpenAI: gpt-5-nano - Anthropic: claude-sonnet-4-20250514 - Gemini: gemini-2.5-flash - Groq: openai/gpt-oss-20b

You will need API keys in these environment variables: OPENAI_API_KEY, ANTHROPIC_API_KEY, GEMINI_API_KEY, GROQ_API_KEY.

To run these examples locally, set a local flag: - Sys.setenv(LLMR_RUN_VIGNETTES = “true”) - or add LLMR_RUN_VIGNETTES=true to ~/.Renviron

OpenAI: gpt-5-nano

library(LLMR)

cfg_openai <- llm_config(
  provider = "openai",
  model    = "gpt-5-nano",
  
)

chat_oai <- chat_session(cfg_openai, system = "Be concise.")
chat_oai$send("Say a warm hello in one short sentence.")
chat_oai$send("Now say it in Esperanto.")

Anthropic: claude-sonnet-4-20250514

cfg_anthropic <- llm_config(
  provider = "anthropic",
  model    = "claude-sonnet-4-20250514",
  max_tokens = 512   # avoid warnings; Anthropic requires max_tokens
)

chat_claude <- chat_session(cfg_anthropic, system = "Be concise.")
chat_claude$send("Name one interesting fact about honey bees.")

Gemini: gemini-2.5-flash

cfg_gemini <- llm_config(
  provider = "gemini",
  model    = "gemini-2.5-flash-lite",
  
)

chat_gem <- chat_session(cfg_gemini, system = "Be concise.")
chat_gem$send("Give me a single-sentence fun fact about volcanoes.")

Groq: openai/gpt-oss-20b

cfg_groq <- llm_config(
  provider = "groq",
  model    = "openai/gpt-oss-20b",
  
)

chat_groq <- chat_session(cfg_groq, system = "Be concise.")
chat_groq$send("Share a short fun fact about octopuses.")

Using the chat history

Chat sessions remember context automatically:

chat_oai$send("What did I ask you to do in my first message?")
# The model can reference the earlier "Say a warm hello" request

Inspect the full conversation

# View all messages
as.data.frame(chat_oai)

# Get summary statistics
summary(chat_oai)

Structured chat in one call (OpenAI example)

schema <- list(
  type = "object",
  properties = list(
    answer     = list(type = "string"),
    confidence = list(type = "number")
  ),
  required = list("answer", "confidence"),
  additionalProperties = FALSE
)

chat_oai$send_structured(
  "Return an answer and a confidence score (0-1) about: Why is the sky blue?",
  schema
)