GLYPH COVERAGE
—Attach a font to checkTYPE TOOLS / 001
Good type.
Real-world fit.
Your font. Your words. Every breakpoint.
Find missing glyphs and awkward wraps before they ship.
THE WORKBENCH
Put your type to the test.
Start with your font, or try the included open-license example.
390 px viewport25.50 px
BREAKPOINT FLAGS
—Run your first fit checkFALLBACK Δ
—Single-line width at 32 pxBreakpoint fit
AWAITING CHECKReal browser measurements at 320, 390, 768, and 1440 px. Includes fallback where your font lacks coverage.
↔
Your copy, across four widths.
Load a font and run a fit check to see line counts, overflow, and fallback differences.Glyph & font details
CMAP + METRICSNo font inspected yet.
Coverage checks Unicode code points in the font’s cmap. A mapped glyph is not proof of correct shaping. Joiners, variation selectors, complex scripts, ligatures, and emoji sequences need visual review. Font metrics below describe the default instance.
Take the spec with you
FONT-FREE EXPORTExport CSS, or a JSON specification with settings and measured results. Restore the JSON later, then reattach your font and paste your copy to check again.
/* Attach a font and run a fit check. */
No font binaries or original copy in either file. You remain responsible for font licensing and permitted web embedding.
THE FIELD GUIDE
Measure first. Then make the call.
Know what is in the file.
Inspect names, units per em, ascender, descender, and available variable axes. Static fonts simply have no axes to adjust. TTF and OTF with TrueType or CFF outlines, plus WOFF 1, are supported when the parser and browser can load them. WOFF2, TTC/OTC, and CFF2 are outside this release.
Check the words you will use.
A missing code point can silently switch to a fallback font. The checker identifies its position in your copy. Coverage and shaping are different: a sequence can use mapped glyphs and still need specialist visual review.
Test more than a type scale.
Fluid sizing interpolates between your two size and viewport endpoints. Four browser-measured trials expose long words, extra lines, and fallback width changes. Results describe this browser and its installed fallback fonts, not every device.
Your fonts stay yours.
Files and copy are processed in memory. There are no uploads, accounts in the tool, analytics, or automatic saves. Reset or reload clears the session. Export contains metadata, settings, code-point positions, and calculations only.
How should I interpret the measurements?
“Overflow” means the measured text extends horizontally beyond the available text width. “Over budget” means the measured line count exceeds your chosen budget. “Fits” only means these two checks passed. Text width uses browser DOM layout at 32 px, with your tracking and variation settings. Positive fallback delta means fallback is wider. Missing characters can already use fallback inside the primary measurement. Base advances are unkerned font units, not a shaped string width.
How does a restored specification work?
A JSON export is a settings recipe and measurement snapshot. Restoring validates the version and inputs, releases the current font, and clears the copy. It never recreates private font files or text. Reattach the same licensed font and paste your original copy to obtain fresh results. The old measurements remain a labeled snapshot until you rerun.
Where do the methods and example font come from?
Font parsing uses the MIT-licensed opentype.js 1.3.4 library (license). The optional Roboto variable example comes from Google Fonts’ source repository and is distributed under the SIL Open Font License. Rendering uses the browser’s FontFace API and font-variation-settings. FontDrop! is another font inspection tool; Utopia explores fluid type scales. Font Fit Lab brings real-copy diagnostics and a restorable spec into one workbench.