Koncile detects tampering in the PDF or scan of a diploma, and documents generated or edited by AI, then returns a fraud score from 0 to 100. Consistency with the school comes next, alongside the official registries.
Create a free account to run the analysis.
Create a free accountFor document volumes or an API, talk to an expert.
The document checker used by insurers, banks and large companies
4.8 / 5
Start with the source: the National Student Clearinghouse for the many US colleges that use it, or the registrar's office of the school. Then check accreditation, because a diploma mill sells real paper for a school that is not recognised. When no registry answers (older degree, foreign or closed school), analyse the document: school, degree title, dates, signatures, layout, then editing and generator traces in the file.
Verifying a degree starts with a question to whoever conferred it. These channels confirm that a degree exists. None of them looks at the file the candidate sent you.
Many US colleges and universities report degrees to the Clearinghouse, and employers order a verification online with the candidate's details and the school's name. The answer confirms the degree, the major and the conferral date when the school participates. Check first that the school is on the participating list.
When the school does not use the Clearinghouse, its registrar answers degree verification requests directly, usually by email or through a form. Expect a delay, and sometimes a fee. A school that cannot be reached at all is information in itself.
A diploma mill issues real looking paper in the name of a school that is not accredited, or that does not exist. The Department of Education's database of accredited institutions and the CHEA directory let you check the school before you check the diploma. An accreditor you cannot find in either is a red flag too.
Older degrees not yet digitised, closed schools, foreign degrees that need a credential evaluation service, and every case where the answer arrives after the hiring decision. That is where the document analysis takes over.
Diplomas have no standard issuing chain: every school has its own paper, layout and signers. So we have no library of issuers here, as we do for invoices or pay stubs, and we say so. What matters is editing traces, the file's fingerprint and the consistency of what is printed.
Most fakes start from a real scanned diploma whose name, title or honors were erased and retyped. K-Detect, our image analysis model, isolates edited areas, flat patches under the text and characters that do not share the font or grain of the rest of the page.
The same model flags documents generated in whole or in part by AI: a rebuilt seal, a signature too sharp, a background too even. The score shows where to look, and every alert comes with its justification.
Every piece of software writes a PDF its own way: object order, numbering, compression. This fingerprint points to the engine that produced the file, even when the metadata was wiped. On ID documents, 1,617 files from a single generator were recognised this way at once, and the method applies to any generated document.
School, degree title, major, honors, signers and conferral date are extracted and cross checked with each other and with the rest of the file: resume, ID, offer letter. A mismatch is not proof, it is a precise question to ask the candidate or the registrar.
Four cases come up again and again, from the clumsy to the hard to see. Two of them leave traces in the file, two only in the background check.
The most common case: an authentic diploma, often a relative's or the candidate's own, scanned and altered. The name, the degree title, the honors or the year are erased and retyped in an image editor or a PDF editor. What gives way: a background broken under the text, a font close to the original but not the same, a field off its baseline, a file saved again after it was created.
Sites sell diplomas built from templates. They all come out of the same engine: same file structure, same fonts, same flaws. That is their weakness. Once the engine's fingerprint is known, the whole batch is recognised, whatever name is printed on it.
A generative AI assistant reproduces a crest, a seal and a signature from a photo, and edits an existing PDF without effort: in an internal test, Claude and ChatGPT edited a PDF invoice with no difficulty. The result fools the eye more often than the analysis: generated areas, paper texture too even, text too regular, a file produced by a tool no registrar uses.
A namesake's authentic diploma passes every file check: only the registry, cross checked with the ID, reveals it. A diploma mill's paper can be clean too. In both cases the page is not the problem: the person or the school is.
A growing number of schools issue a certified electronic diploma alongside the paper one. When you receive one, validate it with the issuer: it beats any analysis.
A PDF or a credential issued by the school or its provider, CeCredential Trust for example, that the graduate shares and that you validate online against the issuer's record. If the validation fails or the link goes to an unknown site, treat the document as unverified.
Most diplomas still arrive as a scan or a phone photo of the paper original. Nothing in that file is signed. The checks then rely on the image and, for a PDF, on its structure and save history.
Without a tool, these three checks stop clumsy fakes in a few minutes. Always ask for the original file, not a screenshot or a rebuilt PDF.
Run a Clearinghouse verification or write to the registrar. A copy of a diploma is a claim; a registry answer is a fact. A candidate who cannot point you to either deserves a follow up question.
Look up the school's accreditation, then the exact degree title in its catalog for that year. Compare the conferral date with the resume and the candidate's age. A title that is close but not exact, or a school found only on its own website, are red flags.
Use our free PDF metadata viewer: creating software, producer and dates. A diploma that went through a PDF editor or an image editor just before it was sent deserves a call to the registrar.
Koncile's fraud detection API gives you a score from 0 to 100 and detailed explanations of manipulation risks, right inside your own tools.
School, degree title, major, honors and conferral date, ready to cross check.
High, medium or low level, with a one sentence explanation.
High risk from 80.
A one line summary ready for your information system.
curl -X POST https://api.koncile.ai/v1/upload_file \ -H "Authorization: Bearer $API_KEY" \ -F "files=@bachelor-diploma.pdf"requests.post("https://api.koncile.ai/v1/upload_file/", headers={"Authorization": f"Bearer {API_KEY}"}, files=[("files", open("bachelor-diploma.pdf", "rb"))]){ "status": "DONE", "General_fields": { "Institution": {"value": "Example State University", "confidence_score": 0.98}, "Degree_title": {"value": "Bachelor of Science", "confidence_score": 0.97} }, "Fraud_detection": { "risk_score": 77, "summary": "Degree title typed over the scanned image", "signals": [ {"name": "Image retouching", "severity": "high"}, {"name": "Fonts and alignment", "severity": "medium"} ] }}Wherever a diploma opens a job, a contract or a seat in a program, a fake costs money and surfaces late. Here is where the checker fits in.
The diploma is checked with the resume and the ID at the offer stage. Doubtful copies go to the registrar or the Clearinghouse, the others do not slow hiring down.
Hundreds of candidates a week, diplomas sent as photos or PDFs. The check runs on receipt, with a score per document and the detail of every test.
Freelancers often upload a degree or a certification at sign up. The checker is called by API in the onboarding flow, with no manual step.
Graduate programs and schools receive diplomas and transcripts at application. Edited files are flagged before the committee meets, not after the semester starts.
A degree title retyped cleanly on a real diploma often leaves no visible error. The fake diploma detector relies on signals that are invisible to the naked eye, in the pixels of the scan and in the structure of the PDF.
| Diploma element | What Koncile checks | What gives an edit away |
|---|---|---|
| Scanned image | Edited areas, flat patches under the text, AI generated parts, with K-Detect | A background broken under the name or the degree title, text laid over the original image |
| Fonts and alignment | Font, weight and baseline of each field, compared with the rest of the page | A name or honors line retyped in a close font, drifting by a few pixels |
| Structure and metadata | Producer software, creation and modification dates, saved versions | A consumer editor, a file modified after creation, a former value left inside the PDF |
| Generator fingerprint | How the file is written, compared with the fake document generators already identified | The engine of a known batch of fakes, even when the metadata was wiped |
| File nature | Scan, photo, native or hybrid PDF, so each case gets the right thresholds | A native PDF with an image of another diploma pasted in |
| Titles and dates | School, degree title, major, honors, signers and conferral date, cross checked with each other | A conferral date that does not fit the program, honors wording the school does not use |
| External sources | Clearinghouse, registrar and accreditation databases, queried by you | A degree the registry does not know, a school no recognised agency accredits |
A diploma ties a name to an education, and it rarely arrives alone in an application. Founded by a former lawyer, Koncile was built with data protection at the core of its architecture: third party SOC 2 audit, GDPR compliance, secure hosting. Documents are encrypted, deleted after analysis and never used to train a third party model.
Ask the source first: a National Student Clearinghouse verification, or the registrar's office of the school. Check the school's accreditation in the Department of Education or CHEA databases. When no registry answers, analyse the document: titles, dates and signers, then editing and generator traces in the file.
Many do for roles where the degree matters, usually through the Clearinghouse, the registrar or a background check provider. It is rarely systematic for every hire, because each verification takes time. Document analysis covers the cases where no registry answers in time.
For a degree, use the Clearinghouse or the registrar. For a professional certification, use the issuing body's online directory, which most certification bodies publish. For a foreign certificate, ask a credential evaluation service. Then check the file you received for edits.
A certified electronic version of a diploma, issued by the school or its provider, that the graduate shares and that you validate online against the issuer's record. If you receive one, validate it. If validation fails, treat the document as unverified.
Compare the copy with an authentic diploma from the same school and the same years: paper, layout, fonts, signers and seal. Check that the degree title and conferral date exist, then look at the file: text laid over a scanned image, or a PDF that went through a consumer editor just before it was sent, are red flags.
An organisation that sells degrees with little or no study, in the name of a school that is not accredited by a recognised agency, or that does not exist. Its paper can look clean. Check the school's accreditation before you check the diploma.
Yes. Generative AI reproduces crests, seals and signatures, and the eye cannot tell. The traces are elsewhere: generated areas and an even background in the pixels, and the structure of the file. That is what Koncile analyses.
It can help, mainly to spot inconsistent dates or titles. It lacks three things: models specialised in images edited or generated by AI, knowledge of the traces each editing tool leaves, and the fingerprints of fake document generators already identified. A language model is only one part of the machine.
It does not see a namesake's authentic diploma presented as the candidate's own, nor a clean document from a diploma mill: only the registry and an accreditation check reveal them. It has no library of issuers, since diplomas have no standard format. On a photo, the original file's metadata is lost and only the image is judged.
It depends on your volume. The price combines a base subscription and a cost per page or per document processed. The plans are detailed on the pricing page.
Yes. Koncile complies with SOC 2 and GDPR, and an independent firm certifies our security and data protection. Your documents are hosted securely with a provider that itself holds SOC 2 and ISO 27001 certifications.
Each document gets its own method: a proprietary ML model for photos, known metadata and forensics for native PDFs, and contextual cross checks.
Inflated balances, deleted lines, a PDF rewritten with software other than the bank's.
See the detection
Net pay changed, contributions that do not match gross pay, a layout that does not match the payroll software.
See the detection
Photo replaced, date of birth or number changed, an MRZ that no longer matches the fields.
See the detection
Edited amounts, wrong VAT, duplicates and made up suppliers.
See the detectionUpload your bank statements, pay stubs or IDs: Koncile extracts the data and flags fakes in seconds.
Talk to an expert
