SERP preview tool

Use this SERP preview tool for pixel width, not a magic character number

Enter the title link you prefer, a display URL, and your preferred meta description. The preview measures every glyph using Arial-compatible advances and shows both source width and a conservative review budget. Character and Unicode code-point totals remain useful receipts, but they do not decide whether the model clips the preview. Sixty-five narrow i characters and sixty-five wide W characters have the same count and radically different widths.

If a source exceeds the selected review budget, only the visual result receives an ellipsis. The complete entered text remains in memory and in the downloaded JSON receipt. Truncation is code-point-safe, so an astral Unicode symbol is not split into a broken surrogate. The receipt also keeps measured source width, displayed width, model date, font, sizes, line heights, result-column width, and both budgets.

The desktop model is explicit and versioned

This version uses google-organic-desktop-model-2026-08-05. It records a 652 CSS pixel organic-result column, a 22 pixel title with 28 pixel line height, and a 14 pixel description with 22 pixel line height. Text advances come from Liberation Sans Regular, which is designed to match Arial metrics and is independently checked in the same core used by CreatorValet’s publishing gate.

The card above is laid out at that same 652 pixel column, so a line break you see is a line break the model produced. A screen too narrow to hold 652 pixels scrolls sideways across the card instead of reflowing it, because narrowing the column would move every break in it. An earlier version drew the card in whatever width the page shell had left over, and a title measured at 97.6 percent of budget could wrap onto a second line directly beside a receipt that said it fit.

The 580 pixel title budget and 920 pixel total description budget are deliberately conservative CreatorValet review thresholds. They are not published Google cutoffs. A description is rendered inside the 652 pixel result column and can wrap; its 920 pixel budget is the measured text advance across its lines rather than a claim that one line is 920 pixels wide. No mobile toggle appears because a second viewport would need its own current browser measurement and independent fixtures.

The title budget is deliberately narrower than the column, which is why a fitted preview always occupies a single line. Two smaller rules keep the measurement matched to the picture. Formatting characters a browser paints at zero width — zero-width space, joiner and non-joiner, word joiner, byte order mark, the bidi controls, variation selectors and the soft hyphen — are charged nothing, because the metric font has no glyph for most of them and would otherwise bill each one the width of a missing-glyph box. And an international domain is shown as its punycode form in both card rows rather than unicode in one and punycode in the other: punycode is what DNS and the request line carry, it is what a browser falls back to whenever a label mixes scripts, and it is the wider of the two spellings, so a width receipt built on it warns early rather than late. Leave the display URL blank and the card says the field is blank instead of quietly inventing a domain.

A SERP preview is an approximation, not a display promise

Google describes title links as automatically generated from several possible sources. The HTML <title> is one input, but the main visible heading, prominent page text, og:title, internal and external link text, and other signals can influence the displayed link. Google can therefore rewrite even a title that fits this model perfectly.

Search snippets are also query-dependent. Google says they are primarily created from page content and may use the meta description when that gives a more useful summary. Device width changes available space, and search features around an organic result can change its shape. This tool answers the narrower and reproducible question: how would the entered strings fit one named desktop font-and-column model?

Pixel receipts and general character counts solve different jobs

A general character counter measures arbitrary text units, including characters, code points, graphemes, words, lines, and encoded bytes. It does not know that one field is a proposed title link or that another is a search description. This route adds the dated organic-result contract, visual context, separate title and description budgets, and a reproducible font oracle.

Likewise, a live SERP checker is another product. It queries current rankings for a keyword, location, language, and device. This preview neither fetches a page nor checks rank. That separation keeps private drafts local and avoids implying that a mock result can predict its position. Use the result to review clarity and likely fit, then validate the actual indexed title and snippet in Search Console after Google recrawls the page.