There are three kinds of product sold into this space and they are regularly confused. Reference managers hold citations and PDFs. Screening and review tools hold searches, decisions and extraction. And a growing set of AI tools summarize and draft. Only the middle one is what a systematic review actually needs, and the question that separates it from the others is whether the product can hold a reason against a record you excluded. This page is about choosing on that basis.
The one question that sorts software for research reviews
Take any candidate and try to record why one specific paper was excluded, then find that reason again from a different screen a week later. Reference managers have no field for it and never will, because their object is a citation rather than a decision. Screening tools are built around exactly that object. Everything else about the choice is comfort by comparison, because this is the field a flow diagram is counted from.
Where a literature review maker sits
Tools that draft prose or summarize papers can genuinely save reading time, and they cannot make a review. They do not hold a search, a screening decision or an extraction, and text they produce is not something you can submit as your own work in any institution we are aware of. Treat them as reading aids that sit beside the record rather than as the thing that keeps it, and check your institution's rules before using one at all.
What to judge academic research software on
Whether two people can screen the same set independently and the tool shows where they disagreed; whether extraction forms are yours to design rather than fixed; whether everything exports in a form you can still read when the subscription ends; and what it costs for the months your review actually runs rather than per year. A review is a project with an end, and software priced like a permanent platform fits badly.
When a spreadsheet is still the honest answer
Below a few hundred records, with one reviewer and no requirement to publish a flow diagram, a spreadsheet and a reference manager genuinely work. That is why the worksheets on this site are free and need no account: the funnel one turns your own record count into the screening hours that tell you whether you have crossed the line where software pays for itself.
Questions people ask about systematic review software
What software do I need for a systematic review?
Something that holds searches, screening decisions with reasons, and extraction. A reference manager cannot do the middle one, which is the field a flow diagram is counted from. Below a few hundred records a spreadsheet is still honest.
Can AI tools do the screening?
They can rank and suggest, and some reviewers use them to prioritize what to read first. The decision and the record of it stay human, and whether you may use them at all is a question for your institution's rules rather than for the vendor.
Is a literature review maker the same as review software?
No. Drafting tools produce text; review software keeps the method. Text produced by a tool is not submittable as your own work, and it holds none of the record you will be asked for.
What should I check about export?
That searches, decisions with reasons, and extracted data all come out in a form you can still read after the subscription ends. A review outlives the project it was funded by, and locked data is the commonest regret.