Package: screenllm 0.1.0

screenllm: LLM-Assisted Title/Abstract Screening for Systematic Reviews

A turn-key workflow for LLM-assisted systematic-review screening. The package ranks a corpus of titles and abstracts with an ensemble of open-source large language models served locally by 'Ollama', then applies the SAFE stopping rule to identify the records a human should screen. Defaults match the four-LLM mean ensemble and the SAFE configuration recommended by Spillias et al. (2026). A companion 'Shiny' app walks the human reviewer through the records above the stopping point. Complementary to the 'AIscreenR' package of Vembye et al. (2025) <doi:10.1037/met0000769>, which targets cloud-hosted 'GPT' models via the 'OpenAI' API; 'screenllm' targets locally-served open-weights ensembles with an integrated stopping rule.

Authors:Scott Spillias [aut, cre], Laura Avila Turriago [aut], Christopher Brown [aut], Ariane Easton [aut], Jack Roberts [aut], Michael Sievers [aut], Steve Swearer [aut], Andrew Taylor [aut], Brigette Wright [aut], Valeriya Komyakova [aut]

screenllm_0.1.0.tar.gz
screenllm_0.1.0.zip(r-4.7-any)screenllm_0.1.0.zip(r-4.6-any)screenllm_0.1.0.zip(r-4.5-any)
screenllm_0.1.0.tgz(r-4.6-any)screenllm_0.1.0.tgz(r-4.5-any)
screenllm_0.1.0.tar.gz(r-4.7-any)screenllm_0.1.0.tar.gz(r-4.6-any)
screenllm_0.1.0.tgz(r-4.6-emscripten)
manual.pdf |manual.html✨
DESCRIPTION |NEWS
card.svg |card.png
screenllm/json (API)

# Install 'screenllm' in R:

install.packages('screenllm', repos = c('https://s-spillias.r-universe.dev', 'https://cloud.r-project.org'))

Bug tracker:https://github.com/s-spillias/screenllm/issues

On CRAN:

4.40 score 35 exports 23 dependencies

Last updated from:cc5678d528. Checks:9 OK. Indexed: yes.

Target	Result	Time
linux-devel	OK	138
source / vignettes	OK	272
linux-release	OK	155
macos-release	OK	156
macos-oldrel	OK	151
windows-devel	OK	90
windows-release	OK	86
windows-oldrel	OK	91
wasm-release	OK	125

Exports:audit_disagreements backend_mock backend_ollama build_prompt check_setup clear_cache custom_ensemble default_ensemble default_ensemble_light define_criteria delete_artefact delete_model detect_gpu estimate_runtime export_report export_worksheet find_duplicates gpu_status install_prereqs launch_app launch_screening_app list_project_artefacts list_projects load_artefact load_toy_cbfm_criteria ollama_catalog ollama_health ollama_installed_models_detail plan_screening pull_model rank_records read_decisions read_records save_artefact summarise_screening

Dependencies:askpass cli curl digest dplyr fs generics glue httr2 jsonlite lifecycle magrittr openssl pillar pkgconfig R6 rlang sys tibble tidyselect utf8 vctrs withr

screenllm quickstart

Rendered fromquickstart.Rmdusingknitr::rmarkdown

Last update: 2026-07-14
Started: 2026-07-13

Help page	Topics
Surface strong LLM-human disagreements for a manual audit	audit_disagreements
Mock backend for tests	backend_mock
Ollama backend	backend_ollama
Build a screening prompt for one record	build_prompt
Verify local setup for 'screenllm'	check_setup
Clear cached LLM scores for a project	clear_cache
Define a custom LLM ensemble	custom_ensemble
The paper's default LLM ensemble	default_ensemble
A "light" ensemble that runs on a laptop	default_ensemble_light
Define screening inclusion criteria	define_criteria
Delete an artefact (or a whole project)	delete_artefact
Delete an Ollama model to free up disk space	delete_model
Detect a usable GPU for Ollama inference	detect_gpu
Estimate wall-clock time for a ranking run	estimate_runtime
Render the screening report as a self-contained HTML document	export_report
Export the human-screening worksheet	export_worksheet
Detect and remove duplicate records	find_duplicates
Live NVIDIA GPU status snapshot	gpu_status
One-stop setup for a fresh machine	install_prereqs
Launch the full end-to-end Shiny app	launch_app
Launch the interactive screening Shiny app	launch_screening_app
Canonical artefact filenames	list_project_artefacts
List existing project names	list_projects
Load an artefact from a project directory	load_artefact
Load the ready-made criteria for the toy CBFM corpus	load_toy_cbfm_criteria
Curated catalog of Ollama models useful for screening	ollama_catalog
Ping the Ollama server	ollama_health
Installed Ollama models with disk-space metadata	ollama_installed_models_detail
Plan the human screening set with the SAFE stopping rule	plan_screening
Pull an Ollama model to the local server	pull_model
Rank a corpus with an LLM ensemble	rank_records
Read completed screening decisions	read_decisions
Read a corpus of records from disk	read_records
Save an artefact into a project directory	save_artefact
Summarise a completed screening session	summarise_screening

Package: screenllm 0.1.0

screenllm: LLM-Assisted Title/Abstract Screening for Systematic Reviews

Citation

Development and contributors

Readme and manuals

Help Manual

Usage by other packages (reverse dependencies)