Fake PAN card detection: spot edits, forgeries and AI fakes
A PAN card is the physical credential for India's Permanent Account Number, the ten-character taxpayer identifier the Income Tax Department assigns to individuals and companies, and it anchors KYC at banks, brokers, and employers. The number obeys a strict grammar, so a card whose number breaks the pattern was fabricated by someone who did not know the format. The other common fraud swaps the photograph or signature on a genuinely issued card to redirect it to a new holder.
Last updated:
Most often edited
portrait, name and father's name, date of birth
Reviewed by
Banks, brokers, mutual funds, and property registrars
Forgery methods seen on a pan card
Method
How it works and what it leaves behind
- Number grammar
- Every PAN satisfies the five-letter, four-digit, one-letter pattern, with the fourth character naming the holder class and the fifth matching the surname initial. Violations of the structure, or a company-class fourth letter on a card carrying an individual's photograph, mean the number was invented.
- Character-level typography
- Substituted digits or letters usually arrive from a different font pass. Width variation, baseline drift, or spacing that changes mid-number exposes a hand-edited identifier even when the surrounding print is clean.
- Channel layout conformance
- Protean, UTIITSL, and e-PAN editions each place fields in known positions with their own typography. A claimed e-PAN showing physical-card layout, or a card mixing both channels' design elements, was assembled from templates.
- Photo and signature patching
- Redirecting a legitimately issued card to a new holder means swapping the photograph or signature, which leaves cut-line edges, resolution steps, or sharpening halos around the patched region.
- Cross-field agreement
- The printed name, father's name, and birth date must cohere with the surname initial encoded in the number's fifth character, a cheap structural check that forgers routinely fail.
A practical checklist for a pan card
1. Photograph the card
Shoot both sides straight on at full resolution so the number's letterforms survive character-level analysis, and avoid cropping the signature or photograph edges.
2. Parse the number
Confirm the five-letter, four-digit, one-letter grammar and that the fourth character fits the holder shown; the fifth character should match the printed surname's first letter.
3. Compare the fields
Read name, father's name, and date of birth together. Internal contradictions, such as a surname initial that disagrees with the printed name, are structural failures no print quality can excuse.
4. Run a forensic check on the file
Number grammar catches invented identifiers, but patched photos and signatures are pixel-level work. TamperCheck runs 300+ forensic checks on the grammar, the typography, and the patched regions, returning a risk score from 0 to 100 with findings tied to specific regions of the image.
What looks wrong on a pan card
A portrait that does not match the page around it
Zoom in on the edge of the portrait. A pasted or generated image often leaves a faint halo, or lighting and texture that do not match the rest of the document.
Name and father's name that are set in a slightly different hand
Zoom in on the name and father's name and compare weight, spacing and edge sharpness with the lines above and below it. A substituted value is rarely typeset exactly like the original, even when the typeface name matches.
Lighting on the portrait that disagrees with the card
Check whether the shadow and glare on the face match the shadow and glare across the rest of the card. A pasted portrait carries the lighting of wherever it came from.
Fonts that are nearly right
Issuers use one tightly specified typeface. Compare digit shapes and letter spacing in the name and number against a genuine specimen of the same edition.
Edges and corners that were rebuilt
Rounded corners, laminate edges and background patterns are hard to recreate. Look for a pattern that breaks, blurs or repeats around the edited area.
Inside the forensic analysis of a pan card
Portrait substitution and face-swap analysis
The document is isolated from whatever it was photographed on, then the portrait region is assessed against the card surface around it. A pasted or generated face rarely matches the substrate it was placed onto.
Date and validity consistency
Issue, expiry, and period dates are checked against one another and against the file's own creation metadata, which catches backdating and validity extension.
Signature forgery analysis
Signature regions are assessed for ink consistency and for how the stroke sits on the page, which distinguishes a genuine pen stroke from a pasted image of one.
And many more checks
The three above are the layers that carry the most weight on a pan card. Every upload runs the full suite of 300+ checks regardless of document class, spanning file structure and metadata, pixel-level forensics, font and text rendering, optical and print characteristics, provenance signals, AI-generation signatures, and many more checks.
What to expect on a genuine pan card before you test it
The identifier reads as five letters, four digits, and one letter: the fourth character encodes the holder class (P for an individual, C for a company, H for a Hindu undivided family), and the fifth is the first letter of the holder's surname or the entity's name. The card adds photograph, signature, name, father's name for individuals, and date of birth or incorporation.
Physical cards come from two channels, Protean (formerly NSDL e-Gov) and UTIITSL, whose layouts differ while carrying the same fields, and the paperless e-PAN arrives as a digitally signed PDF from the same channels. A reissued card keeps the same number, so two cards with different numbers claiming one person is a compliance question rather than a routine replacement.
Pan card fraud: your questions answered
What does the fourth character in a PAN tell you?
It names the holder class: P for an individual, C for a company, H for a Hindu undivided family, and so on. A fourth character that contradicts the card's photograph or name field is an internal inconsistency worth treating as a fabrication signal.
Is a printed e-PAN acceptable for KYC?
Yes, the e-PAN is a legally valid credential with the same ten-character number. It is digitally signed and laid out differently from the physical card, so verifiers should expect a distinct format rather than a scan of the card design.
Why would one person hold two PAN cards?
Holding two active PANs is an offence, but reissues replace a damaged or updated card under the same number. Two cards for one person with different numbers is the combination worth escalating, while a single reissued card is routine.
How does TamperCheck handle PAN cards?
TamperCheck applies the structural grammar of the number, inspects character-level typography for substituted digits, and checks the photograph and signature regions for patching. Analysis is a single REST endpoint call, the image is processed ephemerally and not stored, and findings reference the exact card regions involved.
Where does a forensic review of a pan card start?
On a pan card, the review prioritises a substituted photo; edited name, father's name, date of birth or signature; and the PAN format. After those, the checks that run on every document cover the rest of a pan card.
What is the difference between a fake and an altered pan card?
An altered pan card starts from a genuine one with changed fields such as portrait, so a real number may still exist. A fabricated one comes from a template or an AI model. Altered documents pass number checks, fabricated ones pass visual review, and each leaves different forensic traces on a pan card.
Why does a valid number not prove a pan card is real?
A number on a pan card can be well-formed, or even belong to a real person, while the document carrying it was built around someone else. Format validation confirms structure, not provenance. Comparing the number against name and father's name and the file itself closes the gap.
What does a forensic check on a pan card cost?
Pricing is per document, pay as you go, with no subscription or minimum. Checking a pan card costs the same whether you send one file or thousands.
Check a document
Upload a PDF or a photo of pan card. TamperCheck returns a risk score from 0 to 100 and marks the region that was edited, usually within seconds.
For developers, the API class is pan_card. See the API and MCP documentation.