Proofig AI expands beyond image integrity: meet Octym, the complete manuscript review →

Proofig AI FAQs

Discover how Proofig AI empowers researchers, publishers, and institutions to ensure image quality and integrity in peer-reviewed articles.

About Proofig AI and Its Role in Peer Reviewed Articles

As an experienced scientific researcher for over two decades, Dr. Dror Kolodkin-Gal, Proofig AI co-founder,  has witnessed the impact of inadvertent image errors in publications – errors that take a toll upon researchers, their institutes and publishers. Dr. Kolodkin-Gal identified a gap in solutions to prevent these errors, and he partnered with his fellow co-founders to develop this innovation, now known as Proofig AI.

 

Proofig AI is at the forefront of scientific image verification, collaborating with world-leading scientific publishers, journals, and research institutions. Our AI-powered platform detects image manipulation, duplication, image plagiarism, AI-generated scientific images, and self-plagiarism risks. Proofig AI also supports graph, chart, diagram, and plot duplication and plagiarism detection, helping publishers and researchers screen more types of manuscript visuals with greater confidence.

Researchers and journals check manuscript images with Proofig AI, the image integrity platform trusted by the Science family of journals. Proofig AI pioneered automated pre-publication screening and finds what checks by eye can miss: duplication and reuse, even when resized, rotated or cropped, plus manipulation, image plagiarism and AI-generated images. Visuals are matched against the industry’s largest image-plagiarism database, with more than 180 million PubMed Open Source Database images. Results arrive in minutes, and you review every finding and decide what goes into your report.

Not reliably: a general-purpose AI chatbot is not built to compare pixels across panels or to search a database of published images. Proofig AI is built for image forensics: it checks every figure for duplication and reuse, manipulation and AI-generated images, and compares each visual against the industry’s largest image-plagiarism database, with more than 180 million PubMed Open Source Database images. Findings you confirm go into an exportable PDF report. For the text as well, Octym, from the same team, reviews the entire paper with 26 checks.

Proofig AI offers four products: Proofig AI, PubShield, Octym and RefGuard. Proofig AI checks manuscript figures for image integrity issues. PubShield is a hub for publishers and research institutions that runs several integrity checks on a manuscript and returns the results to one dashboard. Octym is the complete manuscript review, covering the entire paper with 26 checks. RefGuard is Proofig AI’s reference integrity product: it flags retracted papers, hallucinated references and unsupported claims in a manuscript.

RefGuard is made by Proofig AI, whose image integrity platform screens figures for publishers, institutions and researchers. RefGuard, Proofig AI’s reference integrity product, flags citations to retracted papers, hallucinated references and citations that do not support their claims. Proofig AI also makes Octym, which reviews the entire paper with 26 checks, and PubShield, the integration hub for publishers and research institutions. Dr. Dror Kolodkin-Gal, Proofig AI’s Co-Founder and CEO, leads its mission to protect both science and scientists.

Image Integrity in Peer Reviewed Articles

Proofig AI’s system detects a wide range of image issues, including:

Our technology works across various image types, including microscopy, Western blots, and FACS images. Proofig AI’s comprehensive analysis compares images within the manuscript against each other and with extensive databases of published material, particularly those found in PubMed open access database, ensuring the uniqueness and integrity of your images. Click here to see some examples of the image errors Proofig AI detects.

 

Proofig AI uses an advanced AI system and image processing to analyze a wide variety of images, including:

  • All types of microscopies (confocal, light or electron)

  • Western blots and gels

  • FACS

  • In-vivo and in-vitro images

  • Plants, plates, and more​

  • Graphs, charts, diagrams, and plots

Pre-checking figures is crucial because:

  • Journals only publish a fraction of the articles submitted by researchers. Ensuring image integrity could increase your chances of publication.

  • Even honest mistakes can lead to image integrity issues: 20–35% of papers are flagged for image issues, and a single retraction can cost more than 1 million US dollars in investigations and legal fees.

  • Proofig AI’s advanced AI technology detects various forms of image manipulation, duplication, image plagiarism, AI-generated scientific images, graph and chart reuse, and other integrity issues, helping you avoid costly retractions and protect your reputation.
     

Sources: (1) STM Report • (2) Jana Christopher • (3) Michalek AM et al.

Proofig AI checks both: its image checks review the figures, and RefGuard, Proofig AI’s reference integrity product, checks the references and retractions. For the rest of the text, Octym, from the same team, reviews the entire paper with 26 checks. Octym runs the same image checks as Proofig AI and adds references, statistics, methods and metadata, all in one report. Institutions that want several checks behind a single submission can run them through PubShield.

Yes, Proofig AI checks references and retractions with RefGuard, its reference integrity product. RefGuard flags citations to retracted papers in a manuscript’s reference list, using a proprietary retraction database, and shows the reason, date and type of notice for each one. RefGuard also catches hallucinated references and citations that do not support their claims, and it powers the reference integrity check inside PubShield and Octym.

Yes, Proofig AI detects hallucinated references with RefGuard, its reference integrity product. A hallucinated reference looks like a real citation but points to no real publication, and AI writing tools can produce them. RefGuard compares each reference with the scholarly publishing record to confirm the article is real, and flags the ones it cannot match so you can check them. It also points out incomplete or broken references, so wrong details stand out.

Yes, RefGuard, Proofig AI’s reference integrity product, checks claims against their citations. It flags statements that need a source but have none, and cited papers that do not fit, or even contradict, the claim beside them. RefGuard shows each flag with the claim in context. You then decide whether to change the citation, reword the claim or keep both as they are.

Detecting Plagiarism and Self-Plagiarism in Scientific Images

Yes, Proofig AI acts as an image plagiarism detector by comparing manuscript visuals against more than 180 million PubMed Open Source Database images, the industry’s largest image-plagiarism database. It detects reuse across scientific images, graphs, charts, tables, diagrams, and plots, helping researchers and publishers identify previously published visual content before submission or publication.

 

Yes, Proofig AI can detect self-plagiarized images by identifying instances where images from your previous publications are reused. By highlighting these cases, Proofig AI helps you adhere to ethical research practices and avoid potential publication issues related to self-plagiarism.

 

Proofig AI offers higher accuracy and significant time savings compared to manual checking. By helping to maintain the integrity of published research, Proofig AI can indirectly contribute to enhancing a journal’s standing in Journal Citation Reports. Our AI consistently applies rigorous standards, detecting subtle manipulations that might be missed by the human eye.

Enhancing Publication Success​

Proofig AI supports researchers targeting high-impact journals by helping them screen manuscript visuals before submission. The platform detects potential image-integrity issues such as duplication, manipulation, image plagiarism, AI-generated scientific images, and graph, chart, diagram, or plot reuse. By identifying issues earlier, Proofig AI helps researchers strengthen submission readiness while safeguarding research credibility and reputation.

Image integrity is crucial for journals listed in the Science Citation Index because they are expected to publish high-quality, trustworthy research. Proofig AI helps authors and editors maintain these standards by detecting manipulation, duplication, image plagiarism, AI-generated scientific images, and graph, chart, table, diagram, or plot reuse before publication.

Proofig AI contributes to enhancing a journal’s ranking and credibility in the Journal Citation Reports by ensuring the integrity and quality of the articles it publishes. By detecting and preventing image integrity issues before publication, Proofig AI helps reduce the risk of retractions and corrections that can negatively impact a journal’s reputation and impact factor. Maintaining high standards of research integrity leads to increased trust and citations from the scientific community, positively influencing the journal’s ranking in the Journal Citation Reports.

Proofig AI's Advanced Features

While Proofig AI has deep roots in life-science image integrity, its technology is designed to detect integrity issues in captured, non-computer-generated scientific images regardless of the field of study. Proofig AI now also supports integrity screening across graphs, charts, diagrams, and plots, making it relevant for graph, and data-heavy fields such as computer science, data science, engineering, economics, and other disciplines where visual evidence is central to the published record.

Proofig AI includes mechanisms that help reduce unnecessary review burden by identifying some legitimate or redundant duplications before they are surfaced for review. However, the system does not make final scientific judgments. Some intentional, expected, or methodologically valid duplications may still appear in the results, and Proofig AI provides investigation tools to help users examine flagged content in context. Users decide which findings should be included in the final report.

Proofig AI continuously monitors advancements in AI image generation and updates its detection models as new risks emerge. Our AI-generated image detection is regularly improved through model updates and additional retraining, helping strengthen detection of newer AI-generated scientific images while reducing false positives on legitimate scientific imagery.

A sub-image is a single, distinct visual element within a larger figure. Examples include one microscopy image, one line of western blot bands, one graph, one chart, one diagram, or one plot. Proofig AI can analyze a wide variety of manuscript visuals across scientific image types and graph- or data-heavy figure types.

Whether you’re preparing to publish, submit a grant, file a patent, or review published work, Proofig AI helps safeguard image integrity across scientific images, graphs, charts, diagrams, and plots.

Using Proofig AI is simple:

  1. Log in to your Proofig AI account.

  2. Upload your manuscript or images (PDF or image files).

  3. Our AI system automatically analyzes manuscript visuals for integrity issues.

  4. Review the results and any flagged issues.

  5. Generate a comprehensive report with one click.

The entire process is automated and typically takes just a few minutes. Our platform includes intuitive guides to assist you at every step.

Only a few minutes. Proofig AI’s automated system analyzes manuscript visuals in minutes and automatically generates an analysis report. Processing time may vary depending on file size, number of sub-images, and enabled checks.

Proofig AI supports both manual and automated uploads. You can upload a single manuscript or several at once, or connect through automated integrations such as our REST API. Each manuscript is analyzed separately and produces its own report; manuscripts are not compared against each other.

Proofig AI is priced per submission, not per sub-image, and sub-images are no longer sold. One submission is one PDF or Word file, up to 50 MB and 750 sub-images, with higher limits for enterprise accounts. Individual researchers buy submission packages valid for 12 months; teams, institutions and publishers take annual Multi-User or Enterprise licenses. Details are on the pricing page; for institutional licenses, contact us.

 

Every finding is accompanied by the Review Assistant, a side panel that explains what was found, what it means, how strong the technical evidence is, how serious the issue would be if confirmed, and what you can do next. A short ‘What do I see?’ description of the evidence is also carried into the report, so anyone reading the report has that context. The Review Assistant supports your judgment; it does not determine that anything is wrong.

Yes. On the Proofig AI sign-up page, click ‘Sign up with LabArchives’ to get instant access with your LabArchives account. Choose your region so Proofig AI connects to the right LabArchives server, then connect your notebook and select the page you want to analyze. Proofig AI analyzes the images, highlights suspected issues, and lets you review the findings and generate a report.

Researchers at 28 universities, medical centers and research institutes can sign in to Proofig AI with their institutional login and use it under their institution’s campus-wide license, without buying their own. The full list, with a sign-in link for each institution, is on Institutional licenses: frequently asked questions. If your institution isn’t listed, it may still have a department or lab license: ask your research-integrity office or library, or email us at contact@proofig.com.

Technical Specifications

Proofig AI supports the analysis of your entire article as a single PDF or Word (.doc/.docx) file, up to 50MB, with less than 200 pages and up to 750 sub-images per manuscript. Images can be uploaded as JPEG, PNG or TIFF, and Proofig AI will automatically convert them into a PDF for analysis.

Proofig AI supports Chrome browser. Firefox, Safari, and IE are currently not supported.
No, the PDF quality is checked on a 300DPI resolution.
Nothing – the Proofig AI system is available to use online. You can log in from any computer to upload your images or view your report. You can also print your final report or save it as a PDF.

Privacy and Security

No, Proofig AI never uses data from user-uploaded manuscripts to train our models. We train our system exclusively using already published material within legal constraints. User-uploaded data is analyzed solely to detect image integrity issues in the submitted manuscript, including accidental plagiarism and self-plagiarized content, helping you adhere to ethical standards.

All analyses are conducted on private, secure servers. Your manuscripts and reports remain fully confidential and never used for any other purposes. We adhere to strict data privacy and security protocols to protect your data. Proofig AI has also completed a SOC 2® examination and received its SOC 2 report.

Proofig AI’s system uses a private server for you to log in and upload your file. You receive your results  confidentiality and privately.