Integrations
LatteReview v1.4.0 integrates decision models like Jev and pplx-decider for systematic literature screening, comparing their performance and cost against LLMs.
Only you can see your notes.
Screenshot unavailable. Open the experiment ↗
In the maker’s comparison of three decision models across 11,793 articles, Perplexity’s pplx-decider had the highest average ranking accuracy. Cases with uncertain decisions can pass to a reviewer that generates text. The evaluations are maker-run, and the package carries a CC BY-NC-ND licence.
LatteReview checks paper titles and abstracts against your inclusion rules and returns scores, without written explanations. You can have a second reviewer examine uncertain cases. Before relying on that handoff, try it with papers whose correct inclusion decisions you already know.