Frequently Asked Questions

It helps detect AI-fabricated references, verifies references against open academic databases, and formats verified references in your selected citation style. Results are decision support; manually review yellow and red results.

Put each reference on a separate line. Numbered lists are supported. You can also upload a .txt, .docx or .pdf file, or drag and drop the file onto the input area.

Yes. .txt, .docx and .pdf files are supported. For a full article, the tool tries to extract the References/Bibliography section. Scanned or image-only PDFs may not contain selectable text and therefore may not be readable.

Green (Verified): a strong, consistent academic record match was found. Yellow (Unverified): no certain match was found; this does not mean fabricated. Red (Suspicious): strong risk signals such as a wrong/dead DOI or author-year-title conflicts were found. No color is an absolute verdict.

No. Green means the tool found a strong and consistent record match; it is not an absolute guarantee. For critical work, also verify the DOI and publisher or journal page.

No. Red indicates strong conflicts or risk signals. The reference may be mistyped, have a wrong DOI or be affected by incomplete database records. Verify it manually using the DOI, journal page, Google Scholar or DergiPark.

The tool does not mark a reference green unless it finds a sufficiently strong match using the title, authors, year and DOI. Very new publications, national or niche journals, books, theses and conference papers may not appear in major indexes and can therefore be marked yellow even when real.

The score measures how consistently the title, authors, publication year, DOI and academic records match. A score of 88/100 indicates a strong match; it does not mean the reference has an 88% probability of being real. Evaluate it together with the color and warnings.

The tool found a similar academic record but is not confident enough that it is your reference. Compare the title, authors, year and DOI; ignore the suggestion if it is not correct.

The authenticity check compares the reference with academic database records. The format check examines punctuation, author order and fields against the selected style, such as APA or IEEE. A format issue does not mean the reference is fabricated.

Automatically correcting an unverified record could insert the wrong source into your bibliography. Only green verified references are formatted automatically; verify yellow and red results manually first.

The selected style does not change the verification result; it changes the corrected bibliography below. If you upload APA references and select IEEE, the corrected version, copied text and corrected CSV field are produced in IEEE style.

More than 20 styles are supported, including APA, Vancouver, IEEE, TÜBİTAK, TÜSEB, MLA, Chicago, Harvard, AMA, ACS, Nature, Science, Cell, The Lancet, BMJ and PLOS. For custom format, verify references first, enter an example reference or rules, then press Apply custom format.

The CSV includes each reference's status, raw reference, matched title, DOI and, for verified references only, the corrected output in the selected style.

Yes. The tool can search using the title, authors and publication year. However, a DOI is strong verification evidence, so references without a DOI or with incomplete details are more likely to be marked yellow.

Book citations are verified against the Google Books and OpenLibrary catalogs; Turkish national journal articles are verified through TR Dizin, the İSAM Turkish Articles Database and SOBIAD (Social Sciences Citation Index). Web sources (institutional pages, news, press releases) are confirmed by visiting the cited address and comparing the page title, and the access date is carried into the corrected reference. For thesis citations, because SOBIAD also indexes the YÖK National Thesis Center records, the thesis is searched there and, if found, its existence is confirmed; otherwise the tool points you to the YÖK National Thesis Center — a yellow result does not mean the thesis is fabricated.

CrossRef, OpenAlex and Semantic Scholar are used for general article search, PubMed and Europe PMC for biomedical publications, DBLP for computer science, ERIC for education research (education journals without DOIs are missing from the major indexes), TR Dizin (TÜBİTAK ULAKBİM) and SOBIAD (Social Sciences Citation Index — including YÖK National Thesis Center records) for Turkish national journals, and the İSAM Turkish Articles Database for Turkish history/culture/theology articles; references containing an arXiv identifier are resolved through Semantic Scholar. Book citations are searched in the Google Books and OpenLibrary catalogs; web sources are confirmed against the cited page itself. DOI resolution and ISSN/journal verification are also used, with CORE open-access search and publisher page metadata as a last resort.

Temporary network, rate-limit or database outages can occur. Try again after a few minutes. Other databases may still work, but a failed query can reduce verification strength, so check the result manually through its DOI, Google Scholar or DergiPark.

Uploaded files are read in your browser and are not uploaded to our server. Bibliography text is not stored; only if you tick “save this result to my account” is the result report saved to your account (My activity), and you can delete it anytime. Some searches, such as CrossRef, OpenAlex, Europe PMC, Google Books and OpenLibrary, run directly from the browser; PubMed, DBLP, TR Dizin, SOBIAD, ERIC, ISSN/journal verification, web-source confirmation and last-resort CORE searches send the required title, ISSN or address through our server without storing it.

Use the 'Is this result wrong? Send feedback' button on the result card. Including the correct DOI or a short explanation helps us review it.

Guests can run 5 searches per day (up to 5 references per search). Free members can check up to 20 references per day and up to 75 references per month (20 per search); exceeding this requires a Pro membership. Pro usage is unlimited; see the Pricing page for current prices.

Author search finds a researcher's real publications in a specific field from the OpenAlex database and lets you add the ones you pick to your bibliography in one click, in the citation style you want (APA 7, IEEE, Vancouver, TÜBİTAK, TÜSEB and 20+ CSL styles). It has three steps: find and select the right author, enter a topic/field, then check the works you like and format them. You can reach the tool from the top menu or directly at /en/author-search.

Results come directly from the OpenAlex academic database; they are not generated by AI. The listed publications are therefore real records with no risk of fabricated (hallucinated) sources. Still, because different authors can share the same name, it is important to pick the right author by institution, ORCID and publication count.

When you search a name, people sharing that name are listed separately. You can tell them apart by institution (university/institute), ORCID identifier, total publication count and citations. Fill in the 'Institution' field in the search box to narrow the results from the start.

When you type a topic, literature search scans publications in that field from the OpenAlex academic database and surfaces the field's publications-per-year, prominent sub-topics, trend and most-cited sources. It also presents candidate research gaps, concept-intersection signals and a theme map. It is meant for quickly understanding a field, gathering sources and generating ideas for your novelty (gap) statement when starting a thesis or paper. Reach it from the top menu or directly at /en/literature-search.

The tool presents gaps as 'candidates', each with its grounding; it does not give a verdict. Every candidate gap has a label: 'from review' (explicitly stated by a review article, most reliable), 'from data' (a counted signal) or 'inferred' (most speculative). The numbers next to a claim link to supporting sources. A real gap must still be confirmed through systematic review and your own domain expertise; the tool is an accelerator, not an oracle.

The feature requires an account but is free. New members get unlimited searches for the first 2 months; after that, 3 searches per day. Beyond that, a Pro membership is required. Repeating the same search does not consume your quota again.

The reference check looks at whether a source actually EXISTS (catching fabricated/hallucinated references). Citation verification goes one step further: even when the source is real, it checks whether the sentence in your text is actually stated by that source, marking each citation as Full / Partial / Unsupported / Could-not-verify. Reach it from the top menu or at /en/citation-check.

First your bibliography is verified and corrected against academic indexes, then your text is split into atomic claims and each is compared by meaning to the content (abstract or full text) of the source it cites; finally you get a report with a corrected bibliography. An English sentence can be matched to a source in another language at the meaning level. Because it fetches full text and runs AI analysis, it can take 1–5 minutes depending on the number of sources. Results are decision support; confirm important flags against the source's full text.

No. “Could not verify” means the source's open-access text could not be reached (closed access/paywall or metadata only); it does not mean the source is fabricated. Verify such sources manually from the source itself.

Add your ORCID and you'll get a weekly email when your publications receive new citations. Unlike Scholar, every citation is vetted: the citing journal's SJR quartile (Q1–Q4), Web of Science index, UBYT status and explainable publishing-risk signals are shown; self-citations are labeled, and a later retraction triggers a separate notice. Identity is ORCID-based, so there are no name mix-ups. Only indexed, verifiable citations are tracked, so counts may be lower than Scholar.

Enter a topic you want to follow; you'll get a weekly email when new publications appear in that area. Up to 3 topic alerts; managed on the My activity page.

Yes, on three layers: (1) During the bibliography check every DOI is compared against our Retraction Watch–based retraction database, and retracted sources are flagged immediately. (2) If a source in a report saved to your account is retracted LATER, you are notified by email (retraction tracking). (3) With the citation alert set up, you are also notified when an article citing your work is later retracted.

The repaired bibliography is the COMPLETE, in-order list you can paste back into your paper: verified entries are rewritten in your selected style, unverified ones are kept in their original form and listed by number for manual checking. Copy it in one click or download it as Word.

A final check before submitting: bibliography health, retraction status, and the target journal's identity, indexing, quartile and explainable risk signals are combined in one Word report. Enter the target journal name or ISSN in the card below the bibliography-check results.

Enter your paper's title (and optionally the abstract); it makes evidence-based journal suggestions by looking at where similar publications actually appeared. Each suggestion comes with the similar publications that justify it, and you can filter by SJR quartile (Q1–Q4), Web of Science membership or the TÜBİTAK UBYT incentive list.

Enter a journal name or ISSN; its identity is checked against OpenAlex, DOAJ, the official ISSN registry, Scimago (SJR), the Web of Science Master Journal List, TÜBİTAK ULAKBİM UBYT and TR Dizin. Explainable signals include an unregistered ISSN, unverified DOAJ coverage combined with a high APC, low impact with high fees, and sudden publication-volume changes. These signals are not a verdict. The licensed WoS quartile (JIF) is not shown; a Master Journal List link is provided.

TÜBİTAK ULAKBİM's UBYT program pays a per-article incentive to Türkiye-affiliated researchers who publish in journals on its list. Journal Check shows the per-article incentive cap (TRY) if the journal is on the current UBYT list, and clearly states when a journal is not on the list (no incentive is paid). The actual payment depends on author count and eligibility; see cabim.ulakbim.gov.tr.

It adapts an editable Word manuscript's page layout, typography, spacing, columns, heading styles, in-text citations and references to a general Springer Nature, Elsevier, IEEE, Wiley, Taylor & Francis or SAGE submission template. These are general publisher profiles; an exact journal-specific format is available as a separately priced custom service.