Paste a reference. We resolve it against the published record and show you what's wrong. A dropped co-author. A DOI off by one character. A year that drifted between online-first and print. Output comes back in APA, MLA, Chicago, Harvard, AMA, ACS, or any of 10,000+ CSL styles.
Your answer will appear here.
Paste a citation. Three things happen. The agent pulls out what it can — surnames, the year, a slice of the title, the journal, a DOI if you gave one. It queries Crossref first, since that's where DOIs live. No hit? It tries Semantic Scholar. Still nothing? Brave for the long tail. Once a canonical record comes back, a reasoning model walks it against your paste field by field. You see the mismatches. You get a clean rewrite in APA 7, MLA 9, Chicago, Harvard, AMA, ACS, or whichever of the 10,000+ CSL styles you picked.
Proofreading misses this stuff. Initials swap. A journal name gets abbreviated wrong. A page range is off by one. One character in the DOI is mistyped and now you're pointing at a different paper entirely. The year sits halfway between online-first and the print issue. These citations look right. They just don't resolve. The checker resolves them and shows you the diff.
Used an LLM to draft your literature review? Run every reference through the checker before you submit. Hallucinated citations are confident, plausible, and completely fake. The only reliable defense is verification against a real database.
Paste your full reference list one entry at a time (or split it into batches) to confirm every citation resolves. Catches missing DOIs, misspelled authors, and formatting drift between APA 6 and APA 7.
Got a reference from a collaborator with no link? The checker resolves it for you and returns the canonical record so you can confirm the source exists and pull the full text.
Restyling a manuscript from MLA to APA usually surfaces hidden errors: missing volume numbers, wrong title case, mismatched dates. Run each citation through the checker after converting to make sure nothing slipped.
Pre-prints often migrate to peer-reviewed venues. The checker spots when a paper you cited as an arXiv pre-print has been formally published and gives you the updated journal record.
If a citation can't be resolved at any major indexer, that's a red flag worth investigating before relying on it in your argument. Retraction notices that overlap the original DOI also surface in the verification step.
Yes. It's the main reason people use it. LLMs invent citations that read real. Plausible authors, plausible journal, plausible DOI. Since every reference gets resolved against Crossref, Semantic Scholar, and PubMed, the fakes have nowhere to hide. No match, flagged. If an LLM touched your literature review at any point, run the whole list through before you submit.
The corrected output comes back in APA 6 or 7, MLA 8 or 9, Chicago (notes-and-bibliography or author-date), Harvard, AMA, ACS, IEEE, Vancouver, Optica, or any of the 10,000+ styles in the CSL registry. Pick your target in the style picker. That's what you'll get back.
You'll see which databases got tried and what queries went out. If the source genuinely isn't indexed — a working paper, an obscure dissertation, non-English work outside the major aggregators — the checker tells you that instead of guessing. You'll also see near-matches when a typo pushed your citation just outside the relevance threshold. Sometimes that one near-match is the answer.
Both. The checker handles journal articles, books with ISBNs, edited-volume chapters, conference papers, technical reports, theses, and websites. Coverage is best for journal articles (Crossref alone holds 130+ million records) and books with ISBNs from major publishers. Personal blogs and ephemeral pages are a different story. They often aren't indexed anywhere, so eyeballing the URL is faster than asking us to verify.
Yes. Paste them as one block. Line breaks, numbered lists, raw bibliography text — the checker splits the input and verifies each entry on its own. Drop in a whole reference section and you'll get a per-entry verdict back. For lists over 100 entries, batch them. Results stream in progressively that way, and one slow lookup won't hold up the rest.
When the DOI resolves to a record carrying a Crossref-registered retraction notice, you'll see it. The checker links straight to the publisher's notice. Citing a retracted paper can sink a medical literature review, so this matters. The caveat: we don't keep our own retraction database. Coverage depends on the publisher registering the retraction with Crossref. Most do. Some take months.
You start with 10 credits on signup and 3 more each day you log in. A typical citation check is 1–2 credits, depending on how many fallback databases the agent had to try. Top-up packs live on the pricing page. Credits don't expire. There's no subscription.
Different jobs. The Citation Generator builds a fresh citation from a DOI, ISBN, URL, or partial title. You don't have one yet; you want one. The Citation Checker takes one you already have and verifies it. Most people use both. Generate the missing entries first. Then run the whole bibliography through the checker before submission.
Not unless you save them to your Works Cited library yourself. Each check runs in memory and the input is gone the moment the response goes out. We log anonymized usage counts (which style was picked, whether the lookup found a match) but never the citation text itself.
Don't have a citation yet? Start with the Citation Generator. Paste a DOI, ISBN, URL, or partial title and get a fully formatted reference in any style.
Have a sentence but no citation? Find Source searches Crossref, Semantic Scholar, arXiv, OpenAlex, and PubMed to surface peer-reviewed papers that actually support your statement.
If you need to switch your reference list from one style to another (MLA to APA, Harvard to Chicago, BibTeX to APA), drop the whole list into the Citation Converter.
Want the bigger picture on why citation precision matters (for academic integrity, for reproducibility, for surviving peer review)? Read Why Citing Sources Matters.
Verifying a citation is only half the job. The source itself also needs to hold up. Our guides on finding credible sources and what counts as peer-reviewed walk through what to look for.
Verifying the reference list is one piece. Your in-text citations also need to match. Read In-Text Citations vs References to see how the two fit together.
The Citation Checker validates the metadata. To verify that an in-text citation actually supports the surrounding claim, drop the PDF into Verify Source.
Once the citations are clean, the Writing Assistant reviews grammar, tone, and clarity so the prose around your references reads as cleanly as the reference list itself.
Most fake citations come from AI drafts. If you used a model anywhere in the writing process, also run the prose through the AI Detector and the Plagiarism Checker.
AI-powered features require an account. The Citation Generator stays free forever.