LLMRagent: Reproducible Language-Model Agents for Research

An 'R' interface built on 'LLMR' creates large language model (LLM) agents for use as reproducible and governed research instruments. Agents pair a model configuration with optional persona instructions. Stateful exchanges retain conversational memory, and native 'R' functions can serve as tools within declared budgets. Several agents can hold a turn-taking conversation over a shared transcript, and factorial experiments across such designs run in parallel. Model calls and tool activity are captured in a run object, so the research record extends beyond generated text. A study manifest hashes the design and computational apparatus without treating sampled replies as part of its identity, and a hash-sealed archive preserves it alongside transcripts and call records. Tool policies record declared side effects and can limit call counts or result sizes; calls marked for human review pause before execution. Robustness checks assess sensitivity to prompt or model changes, while calibration against human labels corrects estimates drawn from imperfect model labels.

Version: 0.8.0
Depends: R (≥ 4.2)
Imports: LLMR (≥ 0.8.10), R6, tibble, rlang, cli, digest, jsonlite, httr2, callr, utils
Suggests: testthat (≥ 3.0.0), knitr, rmarkdown, future, future.apply, dplyr, withr, R.utils, jsonvalidate, igraph, htmltools
Published: 2026-07-21
DOI: 10.32614/CRAN.package.LLMRagent (may not be active yet)
Author: Ali Sanaei [aut, cre]
Maintainer: Ali Sanaei <sanaei at uchicago.edu>
BugReports: https://github.com/asanaei/LLMRagent/issues
License: MIT + file LICENSE
URL: https://github.com/asanaei/LLMRagent, https://asanaei.github.io/LLMRagent/
NeedsCompilation: no
Materials: README, NEWS
CRAN checks: LLMRagent results

Documentation:

Reference manual: LLMRagent.html , LLMRagent.pdf
Vignettes: A deliberation experiment: group composition and collective decisions (source, R code)
Designed conversations: debates, focus groups, interviews, deliberations (source, R code)
LLMRagent in 10 minutes (source, R code)
Governed tools and human review: letting an agent act, safely (source, R code)
Provenance and archiving: an agent run as evidence (source, R code)
Strong planner, cheap workers: one strong model coordinating many cheap ones (source, R code)
Validity: robustness, calibration, and honest claims (source, R code)
Workflows: small, auditable, resumable pipelines (source, R code)

Downloads:

Package source: LLMRagent_0.8.0.tar.gz
Windows binaries: r-devel: not available, r-release: not available, r-oldrel: not available
macOS binaries: r-release (arm64): LLMRagent_0.8.0.tgz, r-oldrel (arm64): LLMRagent_0.8.0.tgz, r-release (x86_64): LLMRagent_0.8.0.tgz, r-oldrel (x86_64): LLMRagent_0.8.0.tgz

Linking:

Please use the canonical form https://CRAN.R-project.org/package=LLMRagent to link to this page.