about.readcube.com

ReadCube's AI assistant is built for literature review work where every answer must be checkable: you ask in plain language, and it returns answers drawn from your own library and trusted databases, with each claim linked to the exact passage in the source. It is most useful when you need to screen, compare, or extract from a defined body of papers rather than explore a topic from scratch. Three capabilities matter most: cited answers, side-by-side comparison of up to 20 papers, and a per-paper AI Access Indicator that flags whether publisher terms permit AI use.

What the assistant actually does

According to ReadCube's own description of the platform, the AI layer sits on top of a literature management system covering discovery, reading, annotation, and citation. The relevant pieces for research use are:

• Cited answers from your sources. You ask a question in plain language, and the assistant pulls answers from your library and trusted databases. Each answer is linked to the exact passage in the source, so you can verify the claim rather than trust the summary.
• Chat with one paper or a whole library. The scope is yours to set — a single PDF or your entire collection.
• Comparison across up to 20 papers. The assistant surfaces patterns, gaps, and contradictions across the selected set, which is the kind of synthesis that normally takes hours of manual note-taking.
• Natural-language search building. Instead of constructing Boolean strings, you describe what you need and the assistant converts it into structured queries.
• AI Access Indicator. For each paper, the platform flags whether publisher terms permit AI use — a check that matters if you plan to feed published text into an AI workflow.

The underlying corpus is large: ReadCube cites access to 165M+ publications and 182M+ patents for question-answering, and 123M+ indexed papers from the Dimensions research database for discovery.

Where it fits in a real review

A concrete example: you are running a systematic review and have 400 records after deduplication. You can ask the assistant to compare a shortlist of 20 candidate studies on, say, sample size and reported outcomes, and it will lay out where they agree and where they conflict — each point tied to a source passage. You then confirm against the PDFs, which are annotated in the same platform.

For screening, ReadCube describes dual-reviewer workflows at abstract and full-text stages, with conflicts automatically flagged for resolution. For extraction, you define custom fields, and AI Autofill pre-fills answers linked to the source for you to confirm. PRISMA flow diagrams and audit trails are generated automatically and export-ready.

That combination — AI drafting plus human confirmation plus an audit trail — is the mechanism that makes the output defensible in a methods section.

Conditions and limits to check before relying on it

• Verification is still your job. The design links every answer to a passage, but the link tells you where the claim came from, not whether the claim is correct. Treat the citation as a pointer to check, not a seal of approval.
• AI permission varies by paper. The AI Access Indicator exists precisely because publisher terms differ. If a paper is flagged as not cleared, don't assume the assistant's use of it is permitted — check the flag per paper.
• Comparison caps at 20 papers. For larger sets, you'll need to batch comparisons or fall back to manual synthesis.
• Pricing and plan details are not established here. ReadCube's site links to a pricing page and offers a free trial, but the specific terms, limits, and costs are not documented in the material available, so confirm them directly before committing a team.

How to evaluate it for your own work

Run a small test before adopting it: pick 10–15 papers you already know well, ask the assistant a question whose answer you can verify, and check whether the linked passages actually support the stated answer. Then try a 20-paper comparison on a question where you know the contradictions. If the citations hold up on material you understand, you have a reasonable basis for trusting it on material you don't.

ReadCube supports a systematic review from first search through PRISMA reporting inside one platform. Its literature-review tools cover dual-reviewer abstract and full-text screening with automatic conflict flagging, custom data-extraction forms, AI Autofill that pre-fills extraction fields with links to the source for you to confirm, and PRISMA flow diagrams and audit trails that build automatically and export in the required format. This fits teams running a structured review who want screening, extraction, and reporting records in a single system rather than spread across separate tools.

Screening at every stage with dual reviewers

ReadCube runs abstract and full-text screening with independent reviewers. Each reviewer screens separately, and when their decisions disagree, the conflict is flagged automatically so the team can review and resolve it. This keeps the dual-reviewer requirement visible in the workflow instead of relying on manual comparison of spreadsheets.

The practical setup is: assign reviewers to a shared library, have each complete screening independently, then work the flagged conflicts until resolved. The expected result is a screening record where every included or excluded study has two independent decisions and a documented resolution.

Consistent data extraction with custom forms

You can define the exact fields your review needs, then extract across every study using the same form. This is the mechanism that keeps extraction consistent: the form fixes what gets captured, so each study is read against identical questions rather than ad hoc notes.

AI Autofill and source-linked confirmation

AI Autofill answers your extraction questions for you, with each answer linked to the source passage. You confirm the result rather than accept it blindly. The workflow is: define fields, let Autofill pre-fill, then check each answer against its linked source before confirming. The audit value comes from that link — every extracted value traces back to a specific passage.

PRISMA reporting and audit trails

PRISMA flow diagrams and audit trails build automatically from the review work and are export-ready in the required format. Because screening decisions, conflicts, and extraction records live in the same platform, the reporting step draws on the actual review history rather than a reconstruction.

Where the rest of the workflow fits

The review tools sit alongside the platform's broader capabilities, which matter for the search and reading stages:

The AI Assistant answers questions from your library and trusted databases, with each answer linked to the exact passage in the source, and can compare up to 20 papers at once to surface patterns, gaps, and contradictions.

What to check before committing

• Review type fit. The screening, extraction, and PRISMA features are described for systematic-style reviews; confirm they match your protocol's specific requirements.
• AI permissions. The AI Access Indicator exists because publisher terms vary per paper — check flagged restrictions before relying on AI extraction for a given source.
• Team access. Shared libraries keep collaborators on the same evidence; confirm how your team will be set up before screening begins.
• Pricing and plan limits. The site links to a pricing page but the supplied material does not state costs or plan restrictions, so check current terms directly at about.readcube.com/pricing before deciding.

For a team already running dual-reviewer screening and PRISMA reporting, the main draw is consolidation: screening conflicts, extraction forms, source-linked AI answers, and the PRISMA diagram all derive from one review record.

A research team can start with ReadCube by using the free trial offered on about.readcube.com, then choosing a plan through the site's Pricing page. This path suits research-driven companies and academic teams that want to test literature discovery, PDF annotation, AI-assisted answers, and review workflows before committing. The site does not publish plan prices in the material reviewed here, so cost and contract terms need to be confirmed on the Pricing page.

What the free trial lets your team evaluate

The trial is framed around the full platform rather than a single feature. According to the site, the platform covers literature management, an AI assistant, literature monitoring, and literature review in one place. A useful trial means testing each of these against your actual project, not just clicking through menus.

Discovery and library setup

• Search access is described as 123M+ indexed papers from the Dimensions research database, with AI answers drawn from 165M+ publications and 182M+ patents.
• Imports are said to arrive with citations, supplemental data, and full text, so you can check whether metadata quality meets your citation needs.
• Tags, custom fields, and Smart Lists are the tools for keeping a growing library sorted and shareable.

Try this during the trial: import a set of references you already know well and confirm that authors, dates, and full text match your originals. This is the fastest way to judge whether the library will hold up for a real review.

Reading and annotation

Highlighting, underlining, drawing, notes, and sticky comments are all done directly on the PDF. Test this on a paper with complex figures, since annotation on dense PDFs is where tools often disappoint.

AI assistant behavior

The AI assistant answers in plain language and links each answer to the exact passage in the source. It can chat with one paper or a whole library, and compare up to 20 papers at once to surface patterns, gaps, and contradictions. It also converts a plain-language request into structured search queries, so teams can skip Boolean strings.

One detail worth checking early: an AI Access Indicator flags, per paper, whether publisher terms permit AI use. If your institution has strict rules on AI and publisher content, review this indicator during the trial rather than after rollout.

Monitoring new research

You can save a recurring search so new papers on a topic are added to your library, with alerts by email on a daily, weekly, or monthly schedule. The site also says AI reads each new result against your criteria and filters out non-matches. Set up one saved search on a topic you follow closely and judge the relevance of what arrives over a week or two.

Systematic review and PRISMA

For teams running structured reviews, the platform supports dual-reviewer abstract and full-text screening with conflicts flagged automatically, custom data extraction forms, AI Autofill of extraction fields linked to sources, and PRISMA flow diagrams with audit trails that are export-ready. These are the features most worth a hands-on test, because screening and extraction are where review timelines are won or lost.

A practical getting-started sequence

1. Open the trial from about.readcube.com. The homepage invites visitors to try for free and explore options.
2. Pick one live project, not a demo. Choose a review or literature scan your team is actually working on so the trial produces usable evidence.
3. Invite a collaborator and share a library. The site states that shared libraries keep collaborators working from the same evidence. Confirm that permissions and updates behave as your team expects.
4. Run a small screening exercise. Assign two reviewers to the same set of abstracts and check that conflicts are flagged for resolution.
5. Build one extraction form. Define the fields your review needs and test AI Autofill, confirming each pre-filled answer links back to its source.
6. Review the Pricing page. Compare plans against team size and the features you actually used, then decide.

Who this fits, and what to confirm first

The main open items before adoption are price and plan limits, both of which sit on the Pricing page rather than in the public descriptions reviewed here. Teams with strict data-governance requirements should also confirm hosting and access policies before importing sensitive unpublished work.

Bottom line

Start with the free trial, run it against one real review with at least two reviewers, and use the Pricing page to match a plan to what your team actually used. If dual-reviewer screening, sourced AI answers, and PRISMA-ready exports are central to your workflow, the trial will show quickly whether ReadCube replaces several separate tools or adds another one.

ReadCube is a literature management platform that covers the research workflow from discovery through systematic review: organizing libraries, reading and annotating PDFs, AI-assisted Q&A with sourced answers, monitoring new publications, and running dual-reviewer screening with PRISMA-ready exports. It fits researchers and research-driven companies that want one platform instead of separate tools for search, reference management, and review screening. The site offers a free trial and a pricing page, so you can evaluate fit before committing.

Literature management: keeping large libraries usable

ReadCube organizes libraries that scale, using tags, custom fields, and Smart Lists to keep growing collections sorted and shareable. Libraries can be shared so collaborators work from the same evidence base.

• Tags and custom fields let you classify papers by project, method, or any scheme your team uses.
• Smart Lists keep filtered views up to date as new items are added.
• Shared libraries mean collaborators see the latest research rather than stale exports.

This matters most when a team's library outgrows folder structures — for example, a lab tracking hundreds of papers across several concurrent projects.

Reading and annotation: marking up evidence in place

Annotation happens directly on the PDF: highlight, underline, draw, and add notes and sticky comments. Imports arrive with full metadata — citations, supplemental data, and full text — so references are usable without manual cleanup.

For a literature review, this is where evidence capture begins: you read a paper, mark the passages that support or contradict your question, and those marks stay attached to the source.

Discovery and AI assistance: answers you can trace

ReadCube's AI is built for scientific work, with several distinct capabilities:

The discovery layer draws on 123M+ indexed papers from the Dimensions research database, and the answer engine covers 165M+ publications and 182M+ patents, with every citation sourced, linked, and verifiable. The AI Access Indicator is worth checking before you rely on AI analysis of a specific paper, since publisher permissions vary.

Literature monitoring: new research comes to you

You can save a recurring search so new papers on your topic are added to your library automatically. Alerts arrive by email on a daily, weekly, or monthly schedule, and AI reads each new result against your criteria to filter out anything that isn't a real match. This replaces manual re-running of searches to catch newly published work.

Systematic review: from first search to PRISMA

For structured reviews, ReadCube supports each screening stage in one place:

• Dual-reviewer screening at abstract and full-text stages, with independent reviewers and conflicts automatically flagged for the team to resolve.
• Custom data extraction forms — define the exact fields your review needs and extract consistently across every study.
• AI Autofill answers your extraction questions, linked to the source, for you to confirm.
• PRISMA flow diagrams and audit trails build automatically and are export-ready.

A practical example: a two-person review team screens 400 abstracts independently; conflicts surface automatically rather than through manual comparison, and the PRISMA diagram is generated from the screening record instead of being assembled by hand.

How to decide if it fits your workflow

ReadCube is a strong fit if you need library management, PDF annotation, sourced AI answers, and systematic-review screening in one platform — particularly for teams running dual-reviewer reviews. If your needs are limited to reference formatting alone, the broader feature set may exceed what you need. Check the pricing page for current plan details and use the free trial to test the AI answer sourcing and screening workflows against your own library before deciding.

ReadCube is a literature management platform for research-driven companies and research teams. It combines discovering, organizing, reading, annotating, sharing, and citing research literature in one place, and extends into AI-assisted search and full literature-review workflows, from first search through PRISMA reporting. It is a fit if your team needs a shared, scalable library and a structured review process; if you only need a simple personal reference manager, the review and collaboration features go beyond that need.

What the platform covers

The site organizes its tools around four stages of a literature review:

• Literature management — libraries that scale, with tags, custom fields, and Smart Lists to keep growing collections sorted and shareable.
• AI assistant — plain-language questions answered from your library and trusted databases, with each answer linked to the exact passage in the source. You can chat with one paper or your whole library and compare up to 20 papers at once to surface patterns, gaps, and contradictions.
• Literature monitoring — save a recurring search and new matching papers are added to your library, with alerts by email on a daily, weekly, or monthly schedule. AI filters new results against your criteria.
• Literature review — dual-reviewer abstract and full-text screening with conflicts flagged automatically, custom data extraction forms, AI Autofill for extraction fields (linked to source, confirmed by you), and PRISMA flow diagrams and audit trails built automatically and export-ready.

Discovery, reading, and citation

• Search: access to 123M+ indexed papers from the Dimensions research database; the AI assistant draws on 165M+ publications and 182M+ patents, with every citation sourced, linked, and verifiable.
• Reading and annotation: highlight, underline, draw, and add notes and sticky comments directly on the PDF.
• Import: references arrive with citations, supplemental data, and full text.
• Search building: describe what you need in plain language and the AI builds it into structured queries, so you don't have to write Boolean strings.
• AI permissions: an AI Access Indicator flags, per paper, whether publisher terms permit AI use — useful when your organization needs to know what is cleared for AI.

Who it is for

The platform is positioned for research-driven companies and teams rather than individual casual readers. The strongest signals are the collaboration features: shared libraries so collaborators work from the same evidence, dual-reviewer screening, and audit trails for systematic reviews. A team running a structured review with multiple reviewers is the clearest fit.

Trying it and checking cost

The site offers a free trial ("Try for free") and an "Explore options" path. Specific plans and prices are on the pricing page (about.readcube.com/pricing); the available material does not state prices, so check that page before committing. Whether a free tier exists beyond the trial is not specified in the available information.

Quick evaluation checklist

If your work is a multi-reviewer literature review or a company-wide research library, the platform's scope matches that task. Start with the free trial on a real review project, then confirm plan details on the pricing page before deciding.