Identifiez une police depuis une image et trouvez des équivalents gratuits.

01 / Image to shortlist

How the font finder by image workflow works

The image font finder uses a short preparation and comparison process. Each step improves the visual evidence before a candidate reaches your results.

1. UPLOAD / 01

   Add a clear reference

   Choose, drag, or paste a PNG, JPG, or WebP file up to 10 MB. A single word or line with several distinct letters usually gives the strongest evidence.

2. PREPARE / 02

   Make the letters easier to read

   Zoom, rotate, increase contrast, or invert the preview when needed. Focus on one type style at a time so unrelated graphics do not influence the analysis.

3. ANALYZE / 03

   Read the typographic traits

   Cloudflare Workers AI examines visible features such as category, likely weight, width, stroke contrast, construction, spacing, and distinctive character shapes.

4. COMPARE / 04

   Review verified candidates

   The response is checked against the local font catalog, then displayed as ranked Google Font alternatives with previews, licenses, and official sources.

A free font finder from image works best when the sample contains enough structure to compare. Four or more different characters are more informative than a single “I,” “O,” or short monogram because they reveal curves, joins, terminals, counters, and proportions.

02 / Correct the wording

Use the font finder by text to improve comparison

Font Finder AI is primarily visual, but its font finder by text field makes the comparison more useful. If you know the words shown in the image, type or correct them before analysis. The result cards can then render the same phrase, letting you compare the original and each candidate under similar conditions. This makes differences in width, spacing, lowercase forms, numerals, and letters such as “a,” “g,” “R,” or “Q” much easier to spot.

Plain text by itself cannot reveal an unknown typeface because it contains no original letter shapes. Text entry supports the image analysis; it does not replace it. You can continue testing a candidate in the font comparison workspace with your own headline, size, and layout.

03 / Evidence, not certainty

What the Font Finder shows in every result

Each result separates visual similarity from licensing. You receive a font name, relative match score, strong, possible, or weak confidence label, a short reason for the ranking, likely category and weight, and an editable specimen using the detected text. You can save promising families locally, open detailed glyph and language information, or compare several candidates together.

A score is a ranking signal, not the probability that the original uses that exact family. Logos often contain redrawn glyphs, custom spacing, outlines, or lettering created from scratch. Honest uncertainty is especially important in an AI font finder: a visually close free alternative can be useful even when the source typeface cannot be proven. Read more about how the matching methodology works and why the interface avoids unsupported accuracy claims.

04 / Open-font catalog

A Google font finder focused on usable alternatives

This Google font finder is an independent tool, not an official Google product. It compares the image analysis with a curated catalog of fonts available through Google Fonts and verifies every displayed family against known local metadata. That focus prevents the result from becoming a list of unavailable names with no clear next step.

Many catalog families use the SIL Open Font License, while others may use Apache or Ubuntu font licenses. The result page keeps the license, official source, language coverage, weights, styles, and variable axes separate from the visual score. Always open the linked license and confirm it for your intended personal, commercial, web, app, or redistribution use. You can also browse free Google Fonts directly or use the font license checker before publishing.

06 / Better input, better shortlist

Font finder upload image checklist

Use a sharp, front-facing image with strong separation between text and background. Crop around one heading, label, or wordmark; avoid mixing several fonts in one analysis. Screenshots normally preserve cleaner edges than heavily compressed social images. Photos work best when the camera is parallel to the text and perspective distortion is limited.

If the first result is weak, try another part of the same design, add more distinctive characters, correct the visible wording, rotate the image, or increase contrast. No image font finder can recover details that are absent from a tiny or blurred source. For logos and custom wordmarks, use the dedicated logo font finder and treat the output as a visual shortlist rather than guaranteed identification.