Turn included papers into structured evidence rows covering study design, population, outcomes, and limitations, with each extracted field traceable to its source paper.
Editable workflow
Each stage stays visible, so you can revise the scope or evidence choices before generating again.
Start from the included set produced by screening.
Generate structured evidence cards for each paper.
Inspect the assembled evidence table and flag gaps.
Generate the synthesis from the checked evidence only.
each included study gets its own evidence record
failed extractions are marked with reasons
structured rows feeding the cited draft
LitSynth treats AI as support for a review process. You keep control of which papers enter the evidence set and which claims need another check.
Generate one evidence card per included paper.
Review study design, population, outcomes, and limitations per paper.
See extraction failures flagged with reasons instead of omissions.
Carry the checked table into the synthesis draft with citations intact.
Each evidence card covers study design, population and sample, interventions or exposures, primary outcomes, quantitative results where reported, and limitations. The assembled evidence table presents these fields side by side for comparison.
The paper stays in the table with the failure marked and a reason attached. LitSynth does not silently drop failed extractions, because a missing row is a decision a reviewer should make consciously.
Yes. The evidence table is inspectable before synthesis, and the workflow is designed for you to catch extraction problems before they propagate into the draft.
For most rapid and PRISMA-lite reviews, the generated table covers the fields a spreadsheet would, with the added benefit that every row stays linked to its source paper. Formal reviews with registered protocols may still require independent double extraction.
Search papers, screen the evidence, generate a cited draft, and return to earlier decisions whenever the scope needs to change.
Start your review