ISBN Check Digit Calculator

Introduction to ISBN check characters

The final character of an ISBN is not part of the publisher's numbering at all. It is a check character: a redundant symbol computed from every character that precedes it, chosen so that a corrupted copy of the number almost always fails a one-line arithmetic test. ISO 2108:2017 lists it as the fifth and final element of an ISBN, after the GS1 prefix element, the registration group element, the registrant element and the publication element, and specifies that it “is calculated using a modulus 10 algorithm”. The 10-digit ISBN that preceded it used a modulus 11 rule instead, which is the only reason an ISBN can end in the letter X.

This tool does three separate jobs with that arithmetic. It computes the missing final character from a 9-digit or 12-digit body; it verifies a complete ISBN you already have and tells you what the character should have been if it is wrong; and it converts between the 10-digit and 13-digit forms, recomputing the check character from scratch each time instead of pasting one form onto the other. Every result is accompanied by the full weighted-sum table, so the answer is auditable rather than something you have to take on trust.

Why this matters in practice: an ISBN is the join key that connects a publisher's metadata to wholesalers, retail point-of-sale systems, library catalogues and rights registries. A single mistyped digit does not produce a loud failure; it produces a record that silently attaches to the wrong edition, or to nothing at all. The check character is the cheap defence that catches most of those mistakes at the moment of entry, which is exactly when they are cheap to fix.

How to use the three ISBN modes

Choose a mode first, because it changes what the input field expects.

1. Compute the missing check character — paste the ISBN body: 9 characters for ISBN-10 or 12 for ISBN-13. Leave the final character off. The tool returns the character that completes the number.
2. Verify a complete ISBN — paste all 10 or 13 characters. The tool recomputes the check character independently and tells you whether the number you hold is self-consistent. If it is not, you are shown the character that would make it valid, which usually points straight at the typo.
3. Convert between ISBN-10 and ISBN-13 — paste a complete ISBN in either form. A 10-digit number becomes a 978 number; a 978 number becomes a 10-digit number. A 979 number has no 10-digit form, and the tool says so rather than fabricating one.

Set Number form to Detect automatically unless you deliberately want to force one interpretation; the automatic setting reads the length of the cleaned input and picks the right rule. Hyphens, spaces and a leading “ISBN” label are stripped before anything is calculated, and a lowercase x is treated as the value 10, so you can paste straight from a catalogue record, a spreadsheet cell or the back of a cover proof.

After a successful calculation you get four things: the check character on its own, the complete number, a per-digit weighted-sum table showing exactly how the total was reached, and an error-detection audit that tests the finished number against every possible single-character typo and every adjacent swap. Two buttons let you copy the finished ISBN or copy a permalink that reopens the page with the same input already loaded, which is useful when you want to send a colleague the exact case you are looking at.

The general weighted-modulus formula behind both schemes

ISBN-10 and ISBN-13 are both instances of one construction. Take the characters of the identifier, give position $i$ a fixed weight $w_i$, add up the weighted values, and require the total to be congruent to zero modulo some base $m$. ISO/IEC 7064:2003, Information technology — Security techniques — Check character systems, is the standard that catalogues this family of schemes and their error-detection behaviour; the ISBN rules are members of that family rather than exotic one-offs.

The check character is simply the value that makes that congruence true. If the character occupies position $n$ with weight $w_n$, then it is the solution of the congruence below, which exists and is unique whenever $w_n$ is invertible modulo $m$.

Two design choices decide how good the scheme is: the modulus and the weights. A prime modulus makes every non-zero weight invertible, which is what buys ISBN-10 its unusually strong transposition detection. A modulus of 10 is friendlier to hardware and to barcode symbologies, which is what the retail supply chain needed, and it is the reason ISBN-13 gives up a little error-detection power. The sections below take each scheme in turn.

The ISBN-10 modulus-11 formula and the letter X

ISBN-10 uses nine body characters and one check character. The leftmost character is weighted by 10, the next by 9, and so on down to 2, with the check character itself carrying weight 1. Add those products and the total must be a multiple of 11. Because the arithmetic works modulo 11, the check value can come out as 10, and the standard writes that single value as X so the identifier stays exactly ten characters long. RFC 3187, the IETF specification for ISBN-based URNs, describes it in the same terms: “a modulus 11 check digit, using X instead of 10”.

If the computed value is 10, the displayed check character becomes X; anything from 0 to 9 is used directly. Verification uses the equivalent whole-number test below, where the check character contributes with weight 1 and X counts as 10. Note that a second, entirely equivalent phrasing is common in the literature: weight the ten positions 1 through 10 from the left instead of 10 down to 1. Both work, because the two weight vectors sum to $11\sum d_i$ and 11 vanishes modulo 11.

One consequence is worth stating plainly, because it is the single best argument for a prime modulus. Suppose two characters at positions $p$ and $q$ are swapped. The weighted total changes by $(w_p-w_q)(d_p-d_q)$. For ISBN-10 the weight difference is between 1 and 9 and the character difference is between 1 and 10, so neither factor is a multiple of 11; since 11 is prime, their product cannot be either. Every transposition of two characters therefore changes the check character — adjacent or not.

The ISBN-13 modulus-10 formula and its GS1 lineage

ISBN-13 is an EAN-13 / GTIN-13 number that happens to live under the GS1 prefixes 978 and 979, so it uses the GS1 check digit rather than an ISBN-specific one. The GS1 General Specifications, section 7.9.1 “Standard check digit calculations for GS1 data structures”, define one algorithm for every fixed-length numeric GS1 key: weights of 3 and 1 applied alternately from the right, summed, and subtracted from the nearest equal or higher multiple of ten. Read left to right across the twelve body digits of an ISBN-13, that comes out as 1, 3, 1, 3, and so on, which is exactly how Appendix A1.1 of the International ISBN Agency's ISBN Users' Manual states it. Because the arithmetic is modulo 10 the answer is always a single digit from 0 to 9; there is no X in ISBN-13.

The whole-number validity test is the same statement rearranged: the twelve body digits plus the check digit, weighted 1 and 3 alternately, must sum to a multiple of 10.

Conversion between the two forms follows directly. Strip the ISBN-10 check character, prepend the GS1 prefix 978, and recompute the modulus-10 check digit; to go back, strip 978 and the modulus-10 check digit and recompute the modulus-11 one. The nine middle characters are carried across untouched.

Worked example: completing an ISBN-13 and deriving its ISBN-10 twin

Take the number used in Appendix A1.1 of the ISBN Users' Manual, seventh edition: 978-92-95055-12-?. The twelve body digits are 9, 7, 8, 9, 2, 9, 5, 0, 5, 5, 1, 2, and the weights alternate 1, 3, 1, 3 from the left.

9×1 + 7×3 + 8×1 + 9×3 + 2×1 + 9×3 + 5×1 + 0×3 + 5×1 + 5×3 + 1×1 + 2×3 = 126

126 divided by 10 leaves a remainder of 6, and 10 − 6 = 4, so the check digit is 4 and the complete number is 978-92-95055-12-4. That is precisely the value printed in the manual, which makes this a genuine end-to-end test of the implementation rather than a self-referential one. Type 978929505512 into the calculator in compute mode and you will get the same 4, together with the twelve-row weighted-sum table that produced the 126.

Now derive the 10-digit twin. Drop the 978 prefix and the check digit; the nine surviving characters are 929505512. Apply the descending weights 10 down to 2:

9×10 + 2×9 + 9×8 + 5×7 + 0×6 + 5×5 + 5×4 + 1×3 + 2×2 = 267

267 divided by 11 leaves a remainder of 3, and (11 − 3) mod 11 = 8, so the ISBN-10 is 9295055128. Check it the other way round: 267 + 8 = 275 = 11 × 25, a clean multiple of 11. Feed 9295055128 back through convert mode and you return to 9789295055124, closing the loop.

The X case is worth seeing once. The body 964620208 gives 90 + 54 + 32 + 42 + 12 + 0 + 8 + 0 + 16 = 254; 254 leaves a remainder of 1 modulo 11, and (11 − 1) mod 11 = 10, which is written X. The complete number is 964-6202-08-X. X is not a letter code for anything; it is the numeral 10 squeezed into one character position.

Why modulus 11 catches every transposition and modulus 10 does not

This is the most useful thing to understand about the two schemes, and the calculator demonstrates it live on whatever number you enter. Both rules catch every single mistyped character: the change to the weighted total is a weight multiplied by a non-zero difference, and every ISBN weight (1, 2, 3, … 10 for ISBN-10; 1 and 3 for ISBN-13) is coprime to its modulus, so the total cannot land back on the same residue.

Transpositions separate them. Swapping two characters shifts the total by the weight difference times the character difference. Under modulus 11 both factors are smaller than 11 and non-zero, and 11 is prime, so the product is never zero — ISBN-10 detects all of them. Under modulus 10 the adjacent weight difference is 3 − 1 = 2, and 2 × (a − b) is a multiple of 10 exactly when a − b is a multiple of 5. With a ≠ b that means a swap of adjacent digits differing by exactly 5 is invisible to ISBN-13: 16 typed as 61, 27 as 72, 38 as 83, 49 as 94, 05 as 50.

Running that experiment over 20,000 random ISBNs gives 0 undetected adjacent swaps for ISBN-10 and about 9.1% undetected for ISBN-13 — and every single one of the ISBN-13 misses had a digit difference of 5, exactly as the algebra predicts. The result panel shows you the specific vulnerable positions in your number, which is a far more concrete warning than a general statistic. This is the price the book trade paid to move ISBNs into the general EAN/GTIN barcode space, and it was a deliberate trade.

How to interpret the result correctly

A passing check character proves one thing only: the string is arithmetically self-consistent. It does not prove that the registration group, registrant and publication elements were ever allocated, that the number belongs to the title you think it does, or that any book was ever published under it. Roughly one random 13-digit string in ten passes the modulus-10 test by chance, so “valid check digit” is a necessary condition, never a sufficient one.

This tool deliberately does not print a hyphenated ISBN. Hyphen placement is not derivable from the digits: the boundary between the registration group, registrant and publication elements depends on the allocated ranges published by the International ISBN Agency, and those ranges change as blocks are issued. A calculator that guesses “three digits, then one, then five” is right for many English-language numbers and wrong for, say, 978-99921-58-10-7. What the tool does show is the one split that is determinable from the standard: the three-digit GS1 prefix element at the front of every ISBN-13, per ISO 2108:2017 clause 4.2. For real hyphenation, use the Agency's published range file or your registrant records.

Common situations and what to enter

The table below summarizes the most typical use cases so you can match your input to the correct mode quickly.

Limitations of check-digit arithmetic and of this tool

The honest boundaries are worth stating explicitly, because check digits are routinely over-trusted.

• Self-consistency is not existence. The arithmetic cannot tell you whether an ISBN was ever assigned, to whom, or to which product form. Only the registrant's records or the International ISBN Agency can.
• ISBN-13 misses some transpositions. Adjacent digits differing by 5 can be swapped without changing the check digit. The audit panel lists the exact vulnerable positions in your number.
• Two compensating errors defeat both schemes. Check characters are designed against single human slips, not against systematic corruption or deliberate tampering. ISO/IEC 7064 is about error detection, not security.
• No hyphenation is produced. Element boundaries depend on the Agency's allocated ranges, not on arithmetic. Only the three-digit GS1 prefix of an ISBN-13 is derivable from the number itself.
• Prefix validity is only sanity-checked. The tool warns when an ISBN-13 does not begin with 978 or 979, the prefixes currently allocated to ISBN, but it does not check registration-group or registrant ranges.
• ISBN-10 is legacy. Since 1 January 2007 an ISBN is only valid in the supply chain in its 13-digit form. Treat 10-digit output as a compatibility aid for older systems and archival records, not as a number to assign.

A last sanity check that costs nothing: ISBN-13 answers are always one numeric digit, ISBN-10 answers are one numeric digit or X, and a 979 number never converts. If you see something else, the mode or the number form is wrong.

Sources and standards used on this page

Sources. The modulus-10 algorithm, the worked example 978-92-95055-12-4 and the “nearest higher multiple of ten” phrasing are taken from the International ISBN Agency, The International Standard Book Number System: ISBN Users' Manual, Appendix 1 §A1.1 (sixth and seventh editions). The five ISBN elements, the three-digit GS1 prefix element (clause 4.2), the statement that the check digit “is calculated using a modulus 10 algorithm” (clause 4.6) and the removal of the 10-digit calculation method from the current edition come from ISO 2108:2017, Information and documentation — International Standard Book Number (ISBN), fifth edition; the 10-digit rule with X in the check position is specified in the superseded ISO 2108:2005 and earlier editions. The alternating 3/1 weighting is normative in the GS1 General Specifications, section 7.9.1 “Standard check digit calculations for GS1 data structures”. The modulus-11 description with “X instead of 10” is also given in IETF RFC 3187, Using International Standard Book Numbers as Uniform Resource Names, sections 3.1 and 5. That a 979 ISBN-13 has no 10-digit equivalent is stated by the Library of Congress ISBN Converter. The general theory of weighted-modulus check character systems is ISO/IEC 7064:2003, Information technology — Security techniques — Check character systems.

Frequently asked ISBN check digit questions