Language Model Agents in R for AI Workflows and Research.

LLMAgentR

Overview

LLMAgentR is an R package for building Language Model Agents using a modular state graph execution framework. Inspired by LangGraph and LangChain architectures, it supports iterative workflows for research, data analysis, and automation.

Installation

install.packages("LLMAgentR")

Development version

To get the latest features or bug fixes, you can install the development version of LLMAgentR from GitHub:

# If needed
install.packages("remotes")

remotes::install_github("knowusuboaky/LLMAgentR")

See the full function reference or the package website for more details.

Environment Setup

API Setup

Sys.setenv(
  OPENAI_API_KEY     = "your-openai-key",
  GROQ_API_KEY       = "your-groq-key",
  ANTHROPIC_API_KEY  = "your-anthropic-key",
  DEEPSEEK_API_KEY   = "your-deepseek-key",
  DASHSCOPE_API_KEY  = "your-dashscope-key",
  GH_MODELS_TOKEN    = "your-github-models-token"
)

LLM Support (Minimal Wrapper)

The chatLLM package allows you to interact with large language models (LLMs) effortlessly - either through direct calls or via reusable minimal wrappers.

Load the Package

library(chatLLM)

Minimal Wrapper Function

Create a lightweight wrapper around call_llm() for reuse. It optionally provides verbose output:

call_llm(
  prompt     = "Summarize the capital of France.",
  provider   = "groq",
  model      = "llama3-8b",
  temperature = 0.7,
  max_tokens = 200,
  verbose = TRUE
)

my_llm_wrapper <- function(prompt, verbose = FALSE) {
  if (verbose) {
    message("[my_llm_wrapper] Sending prompt to LLM...")
  }

  # Suppress console output but always return the response
  response_text <- if (verbose) {
    call_llm(
      prompt     = prompt,
      provider   = "openai",
      model      = "gpt-4o",
      max_tokens = 3000,
      verbose    = TRUE
    )
  } else {
    suppressMessages(
      suppressWarnings(
        call_llm(
          prompt     = prompt,
          provider   = "openai",
          model      = "gpt-4o",
          max_tokens = 3000,
          verbose    = TRUE
        )
      )
    )
  }

  if (verbose) {
    message("[my_llm_wrapper] Response received.")
  }

  return(response_text)
}

Quick Access Shortcut

Alternatively, preconfigure an LLM call for one-liners:

my_llm_wrapper <- call_llm(
      provider    = "openai",
      model       = "gpt-4o",
      max_tokens  = 3000,
      verbose = TRUE
)

Related Package: chatLLM

The chatLLM package (now available on CRAN) offers a modular interface for interacting with LLM providers including OpenAI, Groq, Anthropic, DeepSeek, DashScope, and GitHub Models.

install.packages("chatLLM")

Agent Articles

Detailed guides now live in pkgdown Articles (one per agent):

• Code Generation Agent
• SQL Query Agent
• Research Agent
• Interpreter Agent
• Document Summarizer Agent
• Data Cleaning Agent
• Forecasting Agent
• Data Wrangling Agent
• Weather Agent
• Feature Engineering Agent
• Visualization Agent

Custom graph workflows:

• Building Custom Agents
• Building Multi-Agent Teams

A full index page is also available:

• Agent Index

License

MIT (c) Kwadwo Daddy Nyame Owusu Boakye.