European Latin · Unicode · type design

Every diacritic mark a European Latin font has to get right.

Browse the diacritics of Europe’s Latin-script languages: how each mark is encoded, which letters carry it, who writes them, and the design conventions and OpenType exceptions behind them.

Position

Language

Case

Type

Checklist

Character
checklist

Font check

Test your font against the atlas.

Drop a font file to see which characters, combining marks and locl language systems it covers. The file is read in your browser and never uploaded.

Drop a font here, or choose a fileOTF, TTF or WOFF. WOFF2 files can be previewed but not checked. Choose font

Official coverage

How many of the letters needed by official national languages, and how many of the others, have a glyph in the font’s character map.

Every language

Which languages have every letter they need, and which letters each one is missing. Legacy, historical and dialectal letters are listed as optional.

Combining marks

Whether the font has the combining marks the atlas letters are built from (acute, caron, comma below and the rest), so text typed as a letter plus a separate mark still displays correctly. The stroke overlays and the turned comma above are listed as optional.

locl language systems

Whether the font registers a locl feature for each language that needs one, such as PLK, ROM or TRK.

Kriszta Szegedi

Author

Kriszta Szegedi

A Budapest-based graphic designer and design manager with a background in typography.

Why I made the Diacritic Atlas

Alongside managing day-to-day design projects, I have been thinking about a project like this for a while. Typography and type design were part of my university studies, and they have stayed with me as a lasting interest ever since.

It was only after I finished my first typeface that I realised how much I still didn’t know about the details and how inconsistent my diacritics were. I wanted to understand it properly, so I went down the rabbit hole and started collecting what I learned in a spreadsheet, just for myself: how each mark is encoded, which languages use it, and above all where the rules break.

Over time the spreadsheet grew into something I kept coming back to, and I realised it would be far more useful shared than sitting on my own computer. With the help of AI, turning it into a website became simple, and the result is this atlas.

I didn’t just want to know how accented letters are put together, but why, and where the rules break.

Feedback and corrections

If you have feedback or spot a mistake in the atlas, write to me on LinkedIn or Instagram. It helps to name the letter or language, and a source if you have one.

LinkedIn Instagram

Glossary

A reference built from the letters up.

The Diacritic Atlas collects the diacritic letters of Europe’s Latin-script orthographies and organises them by mark, so you can see the system behind them: which marks sit above, below or across a letter, how Unicode encodes them, which languages need them, and where a font has to behave differently for one language. The collection may grow to cover more languages later.

Languages are listed by the characters they require, not by how many people speak them. Letters that carry no diacritic, such as ß, æ, œ and þ, are outside the scope of this project. A few letters without a diacritic are included all the same: eth because it is easily confused with Đ, the dotless i and j because accented letters are built on them, and schwa because in Azerbaijani it does the job of a diacritic letter, writing the vowel that Gagauz writes as ä (Azerbaijani itself used ä in 1991–92).

Terms

The terms used across the atlas, from the letters themselves to how Unicode encodes them and how a font handles them.

Letters and marks

Diacritic, marká ş ő: A sign added to a letter that changes its sound, stress or meaning, or turns it into a separate letter of the alphabet. The atlas calls them marks and groups the letters by mark.
Base lettera → á: The letter a mark is added to. Fonts usually build accented letters from the base letter’s glyph and the mark’s glyph.
Precomposed letterá: A letter and its mark encoded together as a single character. Every letter in the atlas is encoded as a single character, so it can be typed and searched as one. Some of them, such as ø, ł, ð and ı, have no decomposition at all: they are letters in their own right, not a base letter plus a mark.
Dotless i and jı ȷ: i and j without their dot. Fonts use them as the base when another mark goes above, because the mark replaces the dot (í, ĵ). In Turkish, Azerbaijani, Crimean Tatar and Gagauz, ı is also a letter of its own.
Case formsé É: Separate versions of the marks for capitals, usually flatter and lower (for example acutecomb.case), so accents fit between the cap height and the line above. This applies to nearly every mark above the letter, so the atlas does not repeat it under each mark. Marks below the letter usually need no case forms.

Unicode

Code pointU+00E1: The number Unicode gives each character, written as U+ followed by a hexadecimal number.
Combining markU+0301: A mark encoded as a character that attaches to the character before it. a followed by U+0301 displays as á.
Spacing mark´ U+00B4: The mark as a standalone character with its own width, used to show or talk about the mark rather than to build letters.
Canonical decompositionU+00E1 = U+0061 + U+0301: Unicode’s official breakdown of a precomposed letter into its base letter and combining marks. Software treats both spellings as the same text.
Unicode blockLatin Extended-A: A named range of code points. Older fonts and keyboard layouts often cover whole blocks, so the block shows where gaps in support tend to appear.

OpenType and font building

Glyph: A drawn shape in the font. One character can have several glyphs (the Polish and the default acute), and one glyph can serve several characters.
Glyph nameaacute · acutecomb.case: The name of a glyph in the font source. The atlas uses the readable names of GlyphData, the naming used by Glyphs and offered in FontLab as Alternative names, such as aacute or scommaaccent. Where no standard name exists, the compiled font stores a production name such as uni0219. Suffixes mark variants of a glyph: .case for capitals, .alt for alternates, .loclPLK for a language form. The comb ending marks a combining mark; it comes from the Adobe glyph lists, where only five combining marks have names in the AGLFN (gravecomb, acutecomb, tildecomb, hookabovecomb, dotbelowcomb). The letter panel shows all three names: the production name, the AGLFN name where there is one, and the Glyphs name.
OpenType featurelocl ccmp mark: A set of rules stored in the font that apps apply to text: they substitute one glyph for another or position glyphs relative to each other.
locl, localized forms: The feature that swaps in language-specific glyphs, such as the Polish kreska or the Romanian comma below. It only works when the text is tagged with its language.
Language system tagPLK ROM TRK: The short OpenType code for a language, such as PLK for Polish or TRK for Turkish. A font’s locl rules are filed under these tags. The Languages view lists each language’s tag.
ccmp: Glyph composition and decomposition. For example it replaces i with dotless ı when a mark above follows (i + U+0301 → í), so the mark does not collide with the dot. Marks below, as in į, leave the dot in place.
mark and mkmk: Mark positioning: mark places marks on base letters, mkmk places marks on other marks, as in Vietnamese or Lithuanian i̇́. Both rely on anchors.
liga, standard ligaturesliga: Joins letter pairs such as fi and ffi into one glyph. In Turkish, Azerbaijani, Crimean Tatar and Gagauz it can hide the dot of i, so letters marked liga flag a ligature conflict.
kern, kerningkern: Spacing adjustments between specific pairs of glyphs. Letters marked kern carry a mark beside the stem (ď ľ ť Ľ) and need extra kerning so it does not collide with what follows.
Stylistic setss01–ss20: A named group of alternate glyphs that users switch on by hand. Letters marked ss have language-specific forms worth offering in a stylistic set too, as a fallback for apps that never apply locl.
Anchor: A point stored in a glyph where a mark attaches, usually one above and one below. Matching anchors on letters and marks let the font combine them correctly.

Official and legacy letters

Two marks on the letter cards show how far a font’s character set has to reach.

Sources

Credits and licence

Diacritic Atlas © 2026 Kriszta Szegedi. Concept, research direction and editing by Kriszta Szegedi. Research, writing and website development assisted by AI (Claude, Anthropic).
The text and data of the Diacritic Atlas are licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 (CC BY-NC-ND 4.0). You may share them for non-commercial purposes if you credit “Diacritic Atlas by Kriszta Szegedi” and link to this site. Commercial use and sharing modified versions are not allowed without permission.
Character names and canonical decompositions are taken from the Unicode Character Database, © Unicode, Inc., used under the Unicode License.
Typefaces: letters in Noto Serif Light and marks in Noto Serif Regular, headings in Noto Serif Display (The Noto Project Authors); Interface in IBM Plex Sans and IBM Plex Mono (IBM) and Noto Sans. All fonts are included in the site as Latin subsets, so the page loads nothing from other servers. All under the SIL Open Font License 1.1.
Glyph names follow GlyphData by Georg Seifert, the naming used by the Glyphs app and offered in FontLab as “Alternative” names. GlyphData © Georg Seifert, used under the BSD 3-Clause License. Glyphs and FontLab are trademarks of their respective owners; the Diacritic Atlas is not affiliated with them. AGLFN names: Adobe Glyph List For New Fonts, © Adobe, used under the BSD 3-Clause License.
Fonts tested with the font check are read in your browser only. They are never uploaded or stored.