<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Usman Bashir</title>
    <description>The latest articles on DEV Community by Usman Bashir (@usman_basheers).</description>
    <link>https://dev.to/usman_basheers</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4065175%2F505f679e-e978-4c9a-9070-6a534301d15f.png</url>
      <title>DEV Community: Usman Bashir</title>
      <link>https://dev.to/usman_basheers</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/usman_basheers"/>
    <language>en</language>
    <item>
      <title>Your AI app design now ships with its design tokens, as a board</title>
      <dc:creator>Usman Bashir</dc:creator>
      <pubDate>Tue, 17 Nov 2026 05:38:14 +0000</pubDate>
      <link>https://dev.to/usman_basheers/your-ai-app-design-now-ships-with-its-design-tokens-as-a-board-1497</link>
      <guid>https://dev.to/usman_basheers/your-ai-app-design-now-ships-with-its-design-tokens-as-a-board-1497</guid>
      <description>&lt;p&gt;Ask an AI for app screens and you get screens. Ask a developer to build them and the first question is: what are the tokens? Which green, which font sizes, which spacing steps?&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;TL;DR:&lt;/strong&gt; when you design app screens or a website with SVG Lab (svglab.app) from ChatGPT or Claude, the project now gets a board called &lt;code&gt;00 Design system&lt;/code&gt;: every colour token with its hex value, every text style, the spacing scale, radii, sizes and the parts (buttons, fields, chips, toggles). SVG Lab builds it from the design tokens the screens use and rebuilds it when they change. Shipped in version 1.65.0 on 13 November 2026.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Feooaz9hngwh4vcuntj61.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Feooaz9hngwh4vcuntj61.png" alt="The 00 Design system board next to the two screens it describes" width="800" height="423"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What is on the board
&lt;/h2&gt;

&lt;p&gt;Everything a developer would otherwise eyedrop out of a mockup:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Colour:&lt;/strong&gt; one swatch per value, labelled with the token names that use it (&lt;code&gt;bg&lt;/code&gt;, &lt;code&gt;surface&lt;/code&gt;, &lt;code&gt;accent&lt;/code&gt;, &lt;code&gt;accent_soft&lt;/code&gt;, &lt;code&gt;success&lt;/code&gt;, &lt;code&gt;danger_text&lt;/code&gt; and so on) and the hex.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Type:&lt;/strong&gt; a table, one row per style (&lt;code&gt;display&lt;/code&gt;, &lt;code&gt;title&lt;/code&gt;, &lt;code&gt;heading&lt;/code&gt;, &lt;code&gt;body&lt;/code&gt;, &lt;code&gt;caption&lt;/code&gt;, &lt;code&gt;label&lt;/code&gt;, &lt;code&gt;button&lt;/code&gt;...), with size / line height, weight, font family and tracking.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Spacing:&lt;/strong&gt; the scale in px on a 4 px grid (&lt;code&gt;2xs&lt;/code&gt; 2 px up to &lt;code&gt;7xl&lt;/code&gt; 80 px).&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Corner radius&lt;/strong&gt;, &lt;strong&gt;control heights&lt;/strong&gt;, &lt;strong&gt;icon sizes&lt;/strong&gt;, &lt;strong&gt;stroke widths&lt;/strong&gt; and screen layout values (screen margin, card padding, touch target, top bar, tab bar).&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Parts:&lt;/strong&gt; primary, secondary, outline and ghost buttons, chips, toggles, checkbox, radio, and text fields including the error state.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;If the design has illustrations, SVG Lab also lists the colours they are drawn in, since those are not tokens but you still need them.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fthp8pp4x87ewvxp6dhge.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fthp8pp4x87ewvxp6dhge.png" alt="The full board" width="800" height="1445"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  It is generated from the tokens, not drawn by the model
&lt;/h2&gt;

&lt;p&gt;This is the part that matters for developers. The board is not a picture the AI drew to look like a style guide. SVG Lab builds it from the same design tokens the layout engine uses for the screens, with the first screen the AI lays out. So:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;The values on the board are the values in the screens.&lt;/li&gt;
&lt;li&gt;The AI makes no extra tool calls for it, so it costs no extra credits.&lt;/li&gt;
&lt;li&gt;When the tokens change, SVG Lab rebuilds the board.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;We checked point 3. Our test app used this token change:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight json"&gt;&lt;code&gt;&lt;span class="p"&gt;{&lt;/span&gt;
  &lt;span class="nl"&gt;"before"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt; &lt;span class="nl"&gt;"accent"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"#2F6B4F"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="nl"&gt;"display"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"Fraunces"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="nl"&gt;"sans"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"DM Sans"&lt;/span&gt; &lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
  &lt;span class="nl"&gt;"after"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;  &lt;span class="p"&gt;{&lt;/span&gt; &lt;span class="nl"&gt;"accent"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"#3B4FD8"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="nl"&gt;"display"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"Bodoni Moda"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="nl"&gt;"sans"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"DM Sans"&lt;/span&gt; &lt;span class="p"&gt;}&lt;/span&gt;
&lt;span class="p"&gt;}&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The board came back with the blue swatches and Bodoni Moda samples in the same step, and SVG Lab worked out new success and danger shades to sit with the new accent.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Ff7y23eol5kbgem8i7bfr.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Ff7y23eol5kbgem8i7bfr.png" alt="Before and after one token change" width="800" height="780"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;A side effect: hand edits on the board get overwritten (the board says so in its header). Change the tokens instead, by asking the AI.&lt;/p&gt;

&lt;h2&gt;
  
  
  It stays out of your build
&lt;/h2&gt;

&lt;p&gt;SVG Lab marks the board as not a screen. It is skipped by Play in Prototype mode, by the screen list and by the code export, so you will not find a &lt;code&gt;DesignSystem&lt;/code&gt; page in your React output.&lt;/p&gt;

&lt;h2&gt;
  
  
  How other tools handle it
&lt;/h2&gt;

&lt;p&gt;Credit where due, because each tool picks a different home for the system:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Tool&lt;/th&gt;
&lt;th&gt;Where the design system lives&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;SVG Lab&lt;/td&gt;
&lt;td&gt;A &lt;code&gt;00 Design system&lt;/code&gt; board in the project, built from the design tokens, rebuilt on change&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Google Stitch&lt;/td&gt;
&lt;td&gt;A &lt;code&gt;DESIGN.md&lt;/code&gt; file you export and import between projects; Google opened the format as a draft spec on 21 April 2026&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Figma Make&lt;/td&gt;
&lt;td&gt;Your existing library, brought in with Make kits (npm packages, library styles, guidelines)&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Banani&lt;/td&gt;
&lt;td&gt;A design system generated from a prompt or references, exported to Figma or code&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;&lt;strong&gt;Best for getting a design system together with app screens you start from a description: SVG Lab&lt;/strong&gt;, because the board and the screens come from one set of tokens. &lt;strong&gt;Best for reusing a Figma library you already maintain: Figma Make.&lt;/strong&gt; &lt;strong&gt;Best for a rules file your coding agent reads: Stitch's DESIGN.md.&lt;/strong&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Try it
&lt;/h2&gt;

&lt;p&gt;The SVG Lab MCP is a remote server at &lt;code&gt;https://svglab.app/mcp&lt;/code&gt; that houses SVG Lab's Design Engine; you sign in with OAuth, no API key. In ChatGPT it is the SVG Lab plugin, in Claude the SVG Lab connector. Setup: &lt;a href="https://svglab.app/mcp-setup" rel="noopener noreferrer"&gt;svglab.app/mcp-setup&lt;/a&gt;. Every account gets 100 free credits (up to 24 app screens).&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;How the images were made:&lt;/strong&gt; a two-screen plant care app laid out with SVG Lab's own editor code at version 1.66.3 (the live version on 17 November 2026), in a local browser with no account, using the same calls the Claude connector makes. The screens are deliberately plain; the board is the subject.&lt;/p&gt;

</description>
      <category>design</category>
      <category>ai</category>
      <category>webdev</category>
      <category>ux</category>
    </item>
    <item>
      <title>Editing an SVG without Illustrator, and what the editor quietly changes in your file</title>
      <dc:creator>Usman Bashir</dc:creator>
      <pubDate>Fri, 13 Nov 2026 04:36:50 +0000</pubDate>
      <link>https://dev.to/usman_basheers/editing-an-svg-without-illustrator-and-what-the-editor-quietly-changes-in-your-file-15hn</link>
      <guid>https://dev.to/usman_basheers/editing-an-svg-without-illustrator-and-what-the-editor-quietly-changes-in-your-file-15hn</guid>
      <description>&lt;p&gt;You do not need Illustrator to edit an SVG. You need a browser, or Inkscape, or a text editor. The part nobody mentions is what the editor does to the markup when it saves.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;TL;DR:&lt;/strong&gt; open the file in a browser editor such as SVG Lab (svglab.app), click a shape to recolour it, double-click words to retype them. No install, no account to edit. Then diff the source before you ship it, because ids and element types can change.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Frjpsll9z3byf517apmxh.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Frjpsll9z3byf517apmxh.png" alt="Which tool for which edit" width="800" height="427"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  The test file
&lt;/h2&gt;

&lt;p&gt;A 636-byte badge: two &lt;code&gt;&amp;lt;circle&amp;gt;&lt;/code&gt;s, one &lt;code&gt;&amp;lt;path&amp;gt;&lt;/code&gt; (a leaf), two &lt;code&gt;&amp;lt;text&amp;gt;&lt;/code&gt; lines, every element with an &lt;code&gt;id&lt;/code&gt;.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fx4rrze7d7itl7ot0d4g6.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fx4rrze7d7itl7ot0d4g6.png" alt="The badge open in SVG Lab" width="800" height="583"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Two edits
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Click the leaf, pick a pink swatch in Fill. &lt;code&gt;#f2c14e&lt;/code&gt; became &lt;code&gt;#ef476f&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;Double-click "Fresh Roast", type "Slow Roast". It stayed a &lt;code&gt;&amp;lt;text&amp;gt;&lt;/code&gt; element.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fh2rtl2fsf7dgsog02sxo.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fh2rtl2fsf7dgsog02sxo.png" alt="The leaf recoloured" width="800" height="583"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fp6j7z9xoosrvjztyd9wo.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fp6j7z9xoosrvjztyd9wo.png" alt="The words retyped" width="800" height="583"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  The diff that matters
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight xml"&gt;&lt;code&gt;&lt;span class="c"&gt;&amp;lt;!-- before --&amp;gt;&lt;/span&gt;
&lt;span class="nt"&gt;&amp;lt;path&lt;/span&gt; &lt;span class="na"&gt;id&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"leaf"&lt;/span&gt; &lt;span class="na"&gt;d&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"M200 95 C250 130 255 190 200 230 C145 190 150 130 200 95 Z"&lt;/span&gt; &lt;span class="na"&gt;fill&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"#f2c14e"&lt;/span&gt;&lt;span class="nt"&gt;/&amp;gt;&lt;/span&gt;
&lt;span class="nt"&gt;&amp;lt;text&lt;/span&gt; &lt;span class="na"&gt;id&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"name"&lt;/span&gt; &lt;span class="na"&gt;x&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"200"&lt;/span&gt; &lt;span class="na"&gt;y&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"290"&lt;/span&gt; &lt;span class="na"&gt;text-anchor&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"middle"&lt;/span&gt; &lt;span class="na"&gt;font-family&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"Georgia, serif"&lt;/span&gt; &lt;span class="na"&gt;font-size&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"40"&lt;/span&gt; &lt;span class="na"&gt;fill&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"#ffffff"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&lt;/span&gt;Fresh Roast&lt;span class="nt"&gt;&amp;lt;/text&amp;gt;&lt;/span&gt;

&lt;span class="c"&gt;&amp;lt;!-- after, from SVG Lab's source panel --&amp;gt;&lt;/span&gt;
&lt;span class="nt"&gt;&amp;lt;path&lt;/span&gt; &lt;span class="na"&gt;d&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"M 200 95 C 250 130 255 190 200 230 C 145 190 150 130 200 95 Z"&lt;/span&gt; &lt;span class="na"&gt;...&lt;/span&gt; &lt;span class="na"&gt;fill&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"#ef476f"&lt;/span&gt;&lt;span class="nt"&gt;/&amp;gt;&lt;/span&gt;
&lt;span class="nt"&gt;&amp;lt;text&lt;/span&gt; &lt;span class="na"&gt;x&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"94.609"&lt;/span&gt; &lt;span class="na"&gt;y&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"260.363"&lt;/span&gt; &lt;span class="na"&gt;font-family&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"Georgia, serif"&lt;/span&gt; &lt;span class="na"&gt;font-size&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"40"&lt;/span&gt; &lt;span class="na"&gt;...&lt;/span&gt; &lt;span class="na"&gt;fill&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"#ffffff"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&amp;lt;tspan&lt;/span&gt; &lt;span class="na"&gt;x&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"94.609"&lt;/span&gt; &lt;span class="na"&gt;dy&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="s"&gt;"0"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&lt;/span&gt;Slow Roast&lt;span class="nt"&gt;&amp;lt;/tspan&amp;gt;&amp;lt;/text&amp;gt;&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Besides my two edits, SVG Lab:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;turned both &lt;code&gt;&amp;lt;circle&amp;gt;&lt;/code&gt;s into &lt;code&gt;&amp;lt;path&amp;gt;&lt;/code&gt;s that draw the same circles,&lt;/li&gt;
&lt;li&gt;dropped every &lt;code&gt;id&lt;/code&gt; (&lt;code&gt;leaf&lt;/code&gt;, &lt;code&gt;name&lt;/code&gt;, &lt;code&gt;tag&lt;/code&gt;),&lt;/li&gt;
&lt;li&gt;swapped &lt;code&gt;text-anchor="middle"&lt;/code&gt; for computed &lt;code&gt;x&lt;/code&gt; positions,&lt;/li&gt;
&lt;li&gt;added a background &lt;code&gt;&amp;lt;rect&amp;gt;&lt;/code&gt; for the artboard colour.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Visually identical. But if your CSS does &lt;code&gt;#leaf { fill: var(--brand) }&lt;/code&gt; or a script grabs &lt;code&gt;#name&lt;/code&gt;, it breaks. Desktop editors add their own markup too (Inkscape writes its namespace unless you save as plain SVG).&lt;/p&gt;

&lt;h2&gt;
  
  
  Which tool when
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Quick colour or text fix, any machine:&lt;/strong&gt; SVG Lab in the browser. Editing is free with no account; downloading needs a free account, and Free includes 3 SVG, PNG or PDF exports a month.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Code-level change:&lt;/strong&gt; SVG Lab's &lt;code&gt;&amp;lt;/&amp;gt;&lt;/code&gt; source panel (edit, Apply) or your own editor.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Long illustration or print work:&lt;/strong&gt; Inkscape, free and open source, currently 1.4.4.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;If the job is a new icon or screen rather than an edit, the SVG Lab MCP lets Claude, ChatGPT, Codex or Cursor draw it as vector shapes through SVG Lab's Design Engine.&lt;/p&gt;

&lt;p&gt;Full walkthrough with FAQ: &lt;a href="https://svglab.app/blog/how-to-open-and-edit-an-svg-without-illustrator" rel="noopener noreferrer"&gt;https://svglab.app/blog/how-to-open-and-edit-an-svg-without-illustrator&lt;/a&gt;&lt;/p&gt;

</description>
      <category>webdev</category>
      <category>svg</category>
      <category>design</category>
      <category>beginners</category>
    </item>
    <item>
      <title>Making a favicon from an SVG, measured: what breaks at 16 pixels</title>
      <dc:creator>Usman Bashir</dc:creator>
      <pubDate>Mon, 09 Nov 2026 04:36:16 +0000</pubDate>
      <link>https://dev.to/usman_basheers/making-a-favicon-from-an-svg-measured-what-breaks-at-16-pixels-16om</link>
      <guid>https://dev.to/usman_basheers/making-a-favicon-from-an-svg-measured-what-breaks-at-16-pixels-16om</guid>
      <description>&lt;p&gt;Every favicon guide gives you the same list of files. I built the set from a real 716-byte logo to see the sizes, and found the one check most guides skip.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;TL;DR:&lt;/strong&gt; a square &lt;code&gt;favicon.svg&lt;/code&gt; (with dark mode inside it), a &lt;code&gt;favicon.ico&lt;/code&gt; with 16, 32 and 48 px images, a 180 px &lt;code&gt;apple-touch-icon.png&lt;/code&gt;, two Android PNGs in a manifest, and four lines in &lt;code&gt;&amp;lt;head&amp;gt;&lt;/code&gt;. Then look at it at 16 px.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F1x5d35gwu2vveso2c1og.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F1x5d35gwu2vveso2c1og.png" alt="The favicon set with measured file sizes" width="800" height="440"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  The files, measured
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;File&lt;/th&gt;
&lt;th&gt;What uses it&lt;/th&gt;
&lt;th&gt;Bytes&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;favicon.svg&lt;/td&gt;
&lt;td&gt;Modern browsers, any size, dark mode&lt;/td&gt;
&lt;td&gt;568 (377 gzipped)&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;favicon.ico&lt;/td&gt;
&lt;td&gt;Old browsers, tools that request /favicon.ico&lt;/td&gt;
&lt;td&gt;2,574&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;apple-touch-icon.png&lt;/td&gt;
&lt;td&gt;iPhone and iPad home screen, 180 x 180&lt;/td&gt;
&lt;td&gt;5,171&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;icon-192.png + icon-512.png&lt;/td&gt;
&lt;td&gt;Android home screen, via the manifest&lt;/td&gt;
&lt;td&gt;20,196&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  1. Make the viewBox square
&lt;/h2&gt;

&lt;p&gt;The logo had a 768 x 896 viewBox. Fitted to the drawing in the SVG Lab editor, the artboard came out 678 x 806. A favicon slot is square, so I set the viewBox to an 860 x 860 square around the mark and dropped the background rectangle.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fhkwzoalxxgnhr8obb4t5.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fhkwzoalxxgnhr8obb4t5.png" alt="The logo in the SVG Lab editor, artboard 678 x 806" width="800" height="500"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  2. Check it at 16 px
&lt;/h2&gt;

&lt;p&gt;I had Chrome render the SVG at 16, 32 and 48 px. At 32 and 48 the bird is clear. At 16 the thin feather tips turn into grey specks. Nothing is wrong with the file: 16 pixels cannot hold a line that thin. If yours fails this check, use a bolder version of the mark.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F31810kfmzf0ggo8sd9rm.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F31810kfmzf0ggo8sd9rm.png" alt="The favicon at 16, 32 and 48 px, enlarged 8 times" width="800" height="427"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  3. Dark mode lives inside the SVG
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight xml"&gt;&lt;code&gt;&lt;span class="nt"&gt;&amp;lt;svg&lt;/span&gt; &lt;span class="na"&gt;xmlns=&lt;/span&gt;&lt;span class="s"&gt;"http://www.w3.org/2000/svg"&lt;/span&gt; &lt;span class="na"&gt;viewBox=&lt;/span&gt;&lt;span class="s"&gt;"59 93 860 860"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&lt;/span&gt;
  &lt;span class="nt"&gt;&amp;lt;style&amp;gt;&lt;/span&gt;
    path { fill: #111 }
    @media (prefers-color-scheme: dark) { path { fill: #f7f3ec } }
  &lt;span class="nt"&gt;&amp;lt;/style&amp;gt;&lt;/span&gt;
  &lt;span class="nt"&gt;&amp;lt;path&lt;/span&gt; &lt;span class="na"&gt;d=&lt;/span&gt;&lt;span class="s"&gt;"M 150 420 C 170 332 ... Z"&lt;/span&gt;&lt;span class="nt"&gt;/&amp;gt;&lt;/span&gt;
&lt;span class="nt"&gt;&amp;lt;/svg&amp;gt;&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Chromium draws SVG favicons in secure static mode, so no web fonts, linked images or scripts: keep everything in the file.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fiazzbaagxiuq33u4aooj.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fiazzbaagxiuq33u4aooj.png" alt="Light and dark renders of the same favicon.svg" width="800" height="347"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  4. The ICO, with Pillow
&lt;/h2&gt;

&lt;p&gt;Chrome rendered the PNGs; Pillow packed the small ones:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight python"&gt;&lt;code&gt;&lt;span class="kn"&gt;from&lt;/span&gt; &lt;span class="n"&gt;PIL&lt;/span&gt; &lt;span class="kn"&gt;import&lt;/span&gt; &lt;span class="n"&gt;Image&lt;/span&gt;
&lt;span class="n"&gt;big&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;Image&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;open&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;fav-48.png&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
&lt;span class="n"&gt;big&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;save&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;favicon.ico&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;sizes&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="p"&gt;[(&lt;/span&gt;&lt;span class="mi"&gt;16&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="mi"&gt;16&lt;/span&gt;&lt;span class="p"&gt;),&lt;/span&gt; &lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="mi"&gt;32&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="mi"&gt;32&lt;/span&gt;&lt;span class="p"&gt;),&lt;/span&gt; &lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="mi"&gt;48&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="mi"&gt;48&lt;/span&gt;&lt;span class="p"&gt;)],&lt;/span&gt;
         &lt;span class="n"&gt;append_images&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="n"&gt;Image&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;open&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;fav-16.png&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;),&lt;/span&gt; &lt;span class="n"&gt;Image&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;open&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;fav-32.png&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;)])&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  5. The head tags
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight html"&gt;&lt;code&gt;&lt;span class="nt"&gt;&amp;lt;link&lt;/span&gt; &lt;span class="na"&gt;rel=&lt;/span&gt;&lt;span class="s"&gt;"icon"&lt;/span&gt; &lt;span class="na"&gt;href=&lt;/span&gt;&lt;span class="s"&gt;"/favicon.ico"&lt;/span&gt; &lt;span class="na"&gt;sizes=&lt;/span&gt;&lt;span class="s"&gt;"32x32"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&lt;/span&gt;
&lt;span class="nt"&gt;&amp;lt;link&lt;/span&gt; &lt;span class="na"&gt;rel=&lt;/span&gt;&lt;span class="s"&gt;"icon"&lt;/span&gt; &lt;span class="na"&gt;href=&lt;/span&gt;&lt;span class="s"&gt;"/favicon.svg"&lt;/span&gt; &lt;span class="na"&gt;type=&lt;/span&gt;&lt;span class="s"&gt;"image/svg+xml"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&lt;/span&gt;
&lt;span class="nt"&gt;&amp;lt;link&lt;/span&gt; &lt;span class="na"&gt;rel=&lt;/span&gt;&lt;span class="s"&gt;"apple-touch-icon"&lt;/span&gt; &lt;span class="na"&gt;href=&lt;/span&gt;&lt;span class="s"&gt;"/apple-touch-icon.png"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&lt;/span&gt;
&lt;span class="nt"&gt;&amp;lt;link&lt;/span&gt; &lt;span class="na"&gt;rel=&lt;/span&gt;&lt;span class="s"&gt;"manifest"&lt;/span&gt; &lt;span class="na"&gt;href=&lt;/span&gt;&lt;span class="s"&gt;"/site.webmanifest"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Note &lt;code&gt;sizes="32x32"&lt;/code&gt;, not &lt;code&gt;sizes="any"&lt;/code&gt;. ChatGPT still gave me &lt;code&gt;any&lt;/code&gt; on 7 November. The Evil Martians favicon guide switched to &lt;code&gt;32x32&lt;/code&gt; in July 2023 to stop Chrome downloading both the ICO and the SVG.&lt;/p&gt;

&lt;p&gt;Browser support: Chrome and Edge from 80, Firefox from 41, Safari from 26 (caniuse, 92.55 percent of browsers in use).&lt;/p&gt;

&lt;p&gt;If you need the icon itself first, the SVG Lab MCP lets Claude, ChatGPT, Codex or Cursor draw icons as vector shapes through SVG Lab's Design Engine.&lt;/p&gt;

&lt;p&gt;Full write-up with FAQ: &lt;a href="https://svglab.app/blog/how-to-make-a-favicon-from-an-svg" rel="noopener noreferrer"&gt;https://svglab.app/blog/how-to-make-a-favicon-from-an-svg&lt;/a&gt;&lt;/p&gt;

</description>
      <category>webdev</category>
      <category>html</category>
      <category>svg</category>
      <category>frontend</category>
    </item>
    <item>
      <title>SVG vs PNG vs WebP, measured on a real logo and a real photo</title>
      <dc:creator>Usman Bashir</dc:creator>
      <pubDate>Sat, 07 Nov 2026 04:37:02 +0000</pubDate>
      <link>https://dev.to/usman_basheers/svg-vs-png-vs-webp-measured-on-a-real-logo-and-a-real-photo-44i7</link>
      <guid>https://dev.to/usman_basheers/svg-vs-png-vs-webp-measured-on-a-real-logo-and-a-real-photo-44i7</guid>
      <description>&lt;p&gt;Every "which image format" answer says the same thing: SVG for logos, WebP for photos, PNG for transparency. It is right, but nobody shows the numbers. So I measured two real images.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;TL;DR:&lt;/strong&gt; SVG for shapes, WebP for anything with a photo in it, PNG only when the pixels must stay exact and the file has to open everywhere.&lt;/p&gt;

&lt;h2&gt;
  
  
  The test
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;A flat bird logo, one path, 716 bytes as SVG. Rendered in headless Chrome at 128, 512 and 1024 px, then saved as PNG and WebP.&lt;/li&gt;
&lt;li&gt;An app screen with an aerial photo map, 739 x 1600, saved as PNG, lossless WebP and WebP at quality 80.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Pillow 12.3 did the encoding: PNG with &lt;code&gt;optimize=True&lt;/code&gt;, WebP with &lt;code&gt;method=6&lt;/code&gt;.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight python"&gt;&lt;code&gt;&lt;span class="kn"&gt;from&lt;/span&gt; &lt;span class="n"&gt;PIL&lt;/span&gt; &lt;span class="kn"&gt;import&lt;/span&gt; &lt;span class="n"&gt;Image&lt;/span&gt;
&lt;span class="n"&gt;im&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;Image&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;open&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;map.png&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;).&lt;/span&gt;&lt;span class="nf"&gt;convert&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;RGB&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
&lt;span class="n"&gt;im&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;save&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;map.webp&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;quality&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="mi"&gt;80&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;method&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="mi"&gt;6&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
&lt;span class="n"&gt;im&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;save&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="s"&gt;map-lossless.webp&lt;/span&gt;&lt;span class="sh"&gt;"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;lossless&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="bp"&gt;True&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;method&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="mi"&gt;6&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fmqlp8ql3hljhney94kp7.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fmqlp8ql3hljhney94kp7.png" alt="File sizes for both test images" width="800" height="360"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Image&lt;/th&gt;
&lt;th&gt;SVG&lt;/th&gt;
&lt;th&gt;PNG&lt;/th&gt;
&lt;th&gt;WebP lossless&lt;/th&gt;
&lt;th&gt;WebP q80&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Logo, 128 px&lt;/td&gt;
&lt;td&gt;716&lt;/td&gt;
&lt;td&gt;4,249&lt;/td&gt;
&lt;td&gt;1,570&lt;/td&gt;
&lt;td&gt;1,582&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Logo, 512 px&lt;/td&gt;
&lt;td&gt;716&lt;/td&gt;
&lt;td&gt;19,075&lt;/td&gt;
&lt;td&gt;6,290&lt;/td&gt;
&lt;td&gt;6,642&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Logo, 1024 px&lt;/td&gt;
&lt;td&gt;716&lt;/td&gt;
&lt;td&gt;41,938&lt;/td&gt;
&lt;td&gt;12,984&lt;/td&gt;
&lt;td&gt;14,416&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Photo screen&lt;/td&gt;
&lt;td&gt;not tested&lt;/td&gt;
&lt;td&gt;1,396,101&lt;/td&gt;
&lt;td&gt;882,060&lt;/td&gt;
&lt;td&gt;139,446&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  What the numbers say
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;SVG does not grow.&lt;/strong&gt; It is 716 bytes at every size, because the browser draws it fresh. The pixel formats grow with the export size. And a small PNG blown up shows its pixels:&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fan9g165v3e1dwbbeinfd.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fan9g165v3e1dwbbeinfd.png" alt="The logo 8 times larger: PNG pixels vs a smooth SVG edge" width="800" height="467"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Lossy WebP is pointless on flat art.&lt;/strong&gt; On the logo, quality 80 was slightly bigger than lossless. Two flat colours leave nothing to throw away.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;On photos, lossy WebP wins by a lot.&lt;/strong&gt; The screen went from 1,396,101 bytes to 139,446, about ten times smaller. The mean pixel difference was 2 levels out of 255.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F0zxiqhfo2hxdxovyr8qi.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F0zxiqhfo2hxdxovyr8qi.png" alt="Zoomed crop: PNG vs WebP at quality 80" width="800" height="467"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Lossless WebP still beats PNG:&lt;/strong&gt; 37 percent smaller on the photo, 69 percent on the logo. PNG earns its place only when a file leaves the web, for example a screenshot someone will edit in a tool that does not read WebP.&lt;/p&gt;

&lt;h2&gt;
  
  
  The rule
&lt;/h2&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Femlm0ofueuxkclobuvye.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Femlm0ofueuxkclobuvye.png" alt="Decision card: SVG for shapes, WebP for photos, PNG for exact pixels" width="800" height="420"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;If your logo only exists as a PNG, trace it back to vector paths first, then tidy the trace with SVGO. The photo in my test is from an app screen a Claude model designed through the SVG Lab MCP, which draws logos, icons and illustrations through SVG Lab's Design Engine as real vector shapes you can export as SVG.&lt;/p&gt;

&lt;p&gt;Full write-up with the FAQ: &lt;a href="https://svglab.app/blog/svg-vs-png-vs-webp" rel="noopener noreferrer"&gt;https://svglab.app/blog/svg-vs-png-vs-webp&lt;/a&gt;&lt;/p&gt;

</description>
      <category>webdev</category>
      <category>performance</category>
      <category>svg</category>
      <category>images</category>
    </item>
    <item>
      <title>I ran SVGO on a traced logo at four precisions. Here is where it breaks.</title>
      <dc:creator>Usman Bashir</dc:creator>
      <pubDate>Thu, 05 Nov 2026 04:37:11 +0000</pubDate>
      <link>https://dev.to/usman_basheers/i-ran-svgo-on-a-traced-logo-at-four-precisions-here-is-where-it-breaks-4d3d</link>
      <guid>https://dev.to/usman_basheers/i-ran-svgo-on-a-traced-logo-at-four-precisions-here-is-where-it-breaks-4d3d</guid>
      <description>&lt;p&gt;Every answer to "how do I make my SVG smaller" says the same thing: run SVGO, lower the precision. Nobody shows the numbers, or says when the precision is too low. So I measured it on a real file.&lt;/p&gt;

&lt;p&gt;The file: a trace of a 716-byte eagle logo, made with VTracer using SVG Lab's Auto-trace settings. The trace came back at 5,647 bytes.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fvdmlegffwgyfvtyibfp3.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fvdmlegffwgyfvtyibfp3.png" alt="The same traced logo at four sizes: 5,647, 2,721, 1,866 and 996 bytes, looking the same" width="800" height="343"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  The command
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;npx svgo logo.svg &lt;span class="nt"&gt;-o&lt;/span&gt; logo.min.svg &lt;span class="nt"&gt;--multipass&lt;/span&gt; &lt;span class="nt"&gt;-p&lt;/span&gt; 2
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;&lt;code&gt;-p&lt;/code&gt; is the precision: how many decimals SVGO keeps. I ran SVGO 4.0.0 with the default preset at 3, 2, 1 and 0.&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Version&lt;/th&gt;
&lt;th&gt;Bytes&lt;/th&gt;
&lt;th&gt;Gzipped&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;As traced&lt;/td&gt;
&lt;td&gt;5,647&lt;/td&gt;
&lt;td&gt;2,531&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Precision 3&lt;/td&gt;
&lt;td&gt;3,703&lt;/td&gt;
&lt;td&gt;1,749&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Precision 2&lt;/td&gt;
&lt;td&gt;2,721&lt;/td&gt;
&lt;td&gt;1,375&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Precision 1&lt;/td&gt;
&lt;td&gt;1,866&lt;/td&gt;
&lt;td&gt;960&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Precision 0&lt;/td&gt;
&lt;td&gt;996&lt;/td&gt;
&lt;td&gt;551&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;h2&gt;
  
  
  What it removed
&lt;/h2&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Ff6522lzqa06vbd40thy3.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Ff6522lzqa06vbd40thy3.png" alt="The first shape before and after SVGO" width="800" height="286"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;The XML header, a generator comment, a version tag, a transform written as &lt;code&gt;253.75001525878906&lt;/code&gt;, and a command letter repeated before every curve. None of it shows on screen.&lt;/p&gt;

&lt;h2&gt;
  
  
  Where precision breaks
&lt;/h2&gt;

&lt;p&gt;Precision counts decimals in the drawing's own units. On a 512-unit logo shown at 512 pixels, one decimal is a tenth of a pixel. On a 24-unit icon the same decimal is far coarser, so do not copy one setting across files.&lt;/p&gt;

&lt;p&gt;I rendered each version at 2x and counted pixels that differ from the trace: 311 at precision 1, 4,078 at precision 0. Zoomed in, precision 0 closes up the thin feather tips.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fm6mu58qlvn3q2cmy2usi.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fm6mu58qlvn3q2cmy2usi.png" alt="6x zoom of the feather tips: precision 0 narrows the gap" width="800" height="343"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;My rule: start at 2, drop to 1 if it still looks right at 400 percent, use 0 only for shapes on whole numbers.&lt;/p&gt;

&lt;h2&gt;
  
  
  Gzip is a second, separate win
&lt;/h2&gt;

&lt;p&gt;Gzip shrinks the download, not the file: 5,647 bytes travel as 2,531. Optimize first, then check the response has &lt;code&gt;Content-Encoding: gzip&lt;/code&gt; or &lt;code&gt;br&lt;/code&gt; in the network tab.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F1eljxthapab7b4iiz47e.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F1eljxthapab7b4iiz47e.png" alt="Raw and gzipped bytes for each version" width="800" height="327"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  The limit of optimizing
&lt;/h2&gt;

&lt;p&gt;The best SVGO result, 996 bytes, is still bigger than the 716-byte hand-made original, and already damaged. Optimizing cleans a trace up; it does not turn it back into a drawing. If the tracer has a detail setting, lower it first (SVG Lab's Auto-trace shows the path count and KB in its preview). If you need a new icon, draw it as vector shapes from the start: the SVG Lab MCP lets the AI you already use draw icons and illustrations through SVG Lab's Design Engine.&lt;/p&gt;

&lt;p&gt;SVG Lab itself has no optimizer today, so for the file, use SVGO or SVGOMG (svgomg.net).&lt;/p&gt;

&lt;p&gt;Full write-up with the FAQ: &lt;a href="https://svglab.app/blog/how-to-make-an-svg-file-smaller" rel="noopener noreferrer"&gt;https://svglab.app/blog/how-to-make-an-svg-file-smaller&lt;/a&gt;&lt;/p&gt;

</description>
      <category>svg</category>
      <category>webdev</category>
      <category>performance</category>
      <category>frontend</category>
    </item>
    <item>
      <title>Curved text in SVG without turning your words into shapes</title>
      <dc:creator>Usman Bashir</dc:creator>
      <pubDate>Sun, 01 Nov 2026 04:37:08 +0000</pubDate>
      <link>https://dev.to/usman_basheers/curved-text-in-svg-without-turning-your-words-into-shapes-5nk</link>
      <guid>https://dev.to/usman_basheers/curved-text-in-svg-without-turning-your-words-into-shapes-5nk</guid>
      <description>&lt;p&gt;Most ways of bending text give you a picture of text. Object to Path in a desktop editor, and the outlined export from many curved text generators, turn every letter into its own shape. It looks right, but you can no longer retype it, swap the font, or let a screen reader read it.&lt;/p&gt;

&lt;p&gt;SVG already has the right tool for this: &lt;code&gt;textPath&lt;/code&gt;.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fm0qjzk8djg39uupe7z5z.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fm0qjzk8djg39uupe7z5z.png" alt="Two panels: converted to paths, where each letter becomes a shape you cannot retype, and real textPath, where the words stay text you can edit, slide and flip" width="800" height="420"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  The markup
&lt;/h2&gt;

&lt;p&gt;Give a path an id, then point a &lt;code&gt;textPath&lt;/code&gt; inside your &lt;code&gt;text&lt;/code&gt; at it:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight xml"&gt;&lt;code&gt;&lt;span class="nt"&gt;&amp;lt;svg&lt;/span&gt; &lt;span class="na"&gt;viewBox=&lt;/span&gt;&lt;span class="s"&gt;"0 0 800 420"&lt;/span&gt; &lt;span class="na"&gt;xmlns=&lt;/span&gt;&lt;span class="s"&gt;"http://www.w3.org/2000/svg"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&lt;/span&gt;
  &lt;span class="nt"&gt;&amp;lt;path&lt;/span&gt; &lt;span class="na"&gt;id=&lt;/span&gt;&lt;span class="s"&gt;"curve"&lt;/span&gt; &lt;span class="na"&gt;d=&lt;/span&gt;&lt;span class="s"&gt;"M 100 320 Q 400 40 700 320"&lt;/span&gt; &lt;span class="na"&gt;fill=&lt;/span&gt;&lt;span class="s"&gt;"none"&lt;/span&gt; &lt;span class="na"&gt;stroke=&lt;/span&gt;&lt;span class="s"&gt;"#7e9bd6"&lt;/span&gt;&lt;span class="nt"&gt;/&amp;gt;&lt;/span&gt;
  &lt;span class="nt"&gt;&amp;lt;text&lt;/span&gt; &lt;span class="na"&gt;font-size=&lt;/span&gt;&lt;span class="s"&gt;"40"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&lt;/span&gt;
    &lt;span class="nt"&gt;&amp;lt;textPath&lt;/span&gt; &lt;span class="na"&gt;href=&lt;/span&gt;&lt;span class="s"&gt;"#curve"&lt;/span&gt; &lt;span class="na"&gt;startOffset=&lt;/span&gt;&lt;span class="s"&gt;"50%"&lt;/span&gt; &lt;span class="na"&gt;text-anchor=&lt;/span&gt;&lt;span class="s"&gt;"middle"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&lt;/span&gt;Text that follows the curve&lt;span class="nt"&gt;&amp;lt;/textPath&amp;gt;&lt;/span&gt;
  &lt;span class="nt"&gt;&amp;lt;/text&amp;gt;&lt;/span&gt;
&lt;span class="nt"&gt;&amp;lt;/svg&amp;gt;&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;ul&gt;
&lt;li&gt;
&lt;code&gt;href&lt;/code&gt; names the path. Older tools want &lt;code&gt;xlink:href&lt;/code&gt; too.&lt;/li&gt;
&lt;li&gt;
&lt;code&gt;startOffset&lt;/code&gt; is where along the curve the words start. &lt;code&gt;50%&lt;/code&gt; with &lt;code&gt;text-anchor="middle"&lt;/code&gt; centres them.&lt;/li&gt;
&lt;li&gt;Text runs the way the path is drawn. To flip it under a curve, reverse the path.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;It is part of SVG 1.1 and renders in Chrome, Edge, Firefox and Safari.&lt;/p&gt;

&lt;h2&gt;
  
  
  Doing it by eye in SVG Lab
&lt;/h2&gt;

&lt;p&gt;Writing the curve by hand is fine for one arc. For anything you want to place by eye, SVG Lab (svglab.app), a free browser SVG editor, binds text to a path in one click. I ran it as a guest, no sign-in, on 31 October 2026. Before, one path and one line of text, both selected:&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fu3cg3dq7fevhd5ios10e.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fu3cg3dq7fevhd5ios10e.png" alt="The SVG Lab artboard with a blue arc and a straight line of text under it, both selected" width="799" height="444"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;After pressing Text on path:&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fpx4v10jlscbb0rvw3bhw.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fpx4v10jlscbb0rvw3bhw.png" alt="The SVG Lab editor with text bent along a blue arc and the On path panel open" width="800" height="279"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;Draw the curve with the Pen or Curve tool, or import an SVG that has one.&lt;/li&gt;
&lt;li&gt;Add the words with the Text tool.&lt;/li&gt;
&lt;li&gt;Select both (Ctrl+A).&lt;/li&gt;
&lt;li&gt;Press &lt;strong&gt;Text on path&lt;/strong&gt; in the Path panel.&lt;/li&gt;
&lt;li&gt;Drag the text to slide it, or use the On path controls: Start, Side (Above or Below), Flip (Upside down) and Length (Natural or Fit path).&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;This is exactly what SVG Lab wrote, copied from its SVG source panel:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight xml"&gt;&lt;code&gt;&lt;span class="nt"&gt;&amp;lt;svg&lt;/span&gt; &lt;span class="na"&gt;xmlns=&lt;/span&gt;&lt;span class="s"&gt;"http://www.w3.org/2000/svg"&lt;/span&gt; &lt;span class="na"&gt;xmlns:xlink=&lt;/span&gt;&lt;span class="s"&gt;"http://www.w3.org/1999/xlink"&lt;/span&gt; &lt;span class="na"&gt;viewBox=&lt;/span&gt;&lt;span class="s"&gt;"0 0 800 420"&lt;/span&gt; &lt;span class="na"&gt;width=&lt;/span&gt;&lt;span class="s"&gt;"800"&lt;/span&gt; &lt;span class="na"&gt;height=&lt;/span&gt;&lt;span class="s"&gt;"420"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&lt;/span&gt;
  &lt;span class="nt"&gt;&amp;lt;rect&lt;/span&gt; &lt;span class="na"&gt;x=&lt;/span&gt;&lt;span class="s"&gt;"0"&lt;/span&gt; &lt;span class="na"&gt;y=&lt;/span&gt;&lt;span class="s"&gt;"0"&lt;/span&gt; &lt;span class="na"&gt;width=&lt;/span&gt;&lt;span class="s"&gt;"800"&lt;/span&gt; &lt;span class="na"&gt;height=&lt;/span&gt;&lt;span class="s"&gt;"420"&lt;/span&gt; &lt;span class="na"&gt;fill=&lt;/span&gt;&lt;span class="s"&gt;"#f7f3ec"&lt;/span&gt;&lt;span class="nt"&gt;/&amp;gt;&lt;/span&gt;
  &lt;span class="nt"&gt;&amp;lt;path&lt;/span&gt; &lt;span class="na"&gt;id=&lt;/span&gt;&lt;span class="s"&gt;"path_path1"&lt;/span&gt; &lt;span class="na"&gt;d=&lt;/span&gt;&lt;span class="s"&gt;"M 100 320 C 300 133.33 500 133.33 700 320"&lt;/span&gt; &lt;span class="na"&gt;stroke=&lt;/span&gt;&lt;span class="s"&gt;"#7e9bd6"&lt;/span&gt; &lt;span class="na"&gt;stroke-width=&lt;/span&gt;&lt;span class="s"&gt;"4"&lt;/span&gt; &lt;span class="na"&gt;stroke-linecap=&lt;/span&gt;&lt;span class="s"&gt;"butt"&lt;/span&gt; &lt;span class="na"&gt;stroke-linejoin=&lt;/span&gt;&lt;span class="s"&gt;"miter"&lt;/span&gt; &lt;span class="na"&gt;fill=&lt;/span&gt;&lt;span class="s"&gt;"none"&lt;/span&gt;&lt;span class="nt"&gt;/&amp;gt;&lt;/span&gt;
  &lt;span class="nt"&gt;&amp;lt;text&lt;/span&gt; &lt;span class="na"&gt;font-family=&lt;/span&gt;&lt;span class="s"&gt;"Arial"&lt;/span&gt; &lt;span class="na"&gt;font-size=&lt;/span&gt;&lt;span class="s"&gt;"40"&lt;/span&gt; &lt;span class="na"&gt;font-weight=&lt;/span&gt;&lt;span class="s"&gt;"400"&lt;/span&gt; &lt;span class="na"&gt;font-style=&lt;/span&gt;&lt;span class="s"&gt;"normal"&lt;/span&gt; &lt;span class="na"&gt;text-anchor=&lt;/span&gt;&lt;span class="s"&gt;"start"&lt;/span&gt; &lt;span class="na"&gt;dominant-baseline=&lt;/span&gt;&lt;span class="s"&gt;"hanging"&lt;/span&gt; &lt;span class="na"&gt;paint-order=&lt;/span&gt;&lt;span class="s"&gt;"fill stroke"&lt;/span&gt; &lt;span class="na"&gt;xml:space=&lt;/span&gt;&lt;span class="s"&gt;"preserve"&lt;/span&gt; &lt;span class="na"&gt;stroke=&lt;/span&gt;&lt;span class="s"&gt;"none"&lt;/span&gt; &lt;span class="na"&gt;stroke-width=&lt;/span&gt;&lt;span class="s"&gt;"2"&lt;/span&gt; &lt;span class="na"&gt;stroke-linecap=&lt;/span&gt;&lt;span class="s"&gt;"round"&lt;/span&gt; &lt;span class="na"&gt;stroke-linejoin=&lt;/span&gt;&lt;span class="s"&gt;"round"&lt;/span&gt; &lt;span class="na"&gt;fill=&lt;/span&gt;&lt;span class="s"&gt;"#1d1d1d"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&amp;lt;textPath&lt;/span&gt; &lt;span class="na"&gt;href=&lt;/span&gt;&lt;span class="s"&gt;"#path_path1"&lt;/span&gt; &lt;span class="na"&gt;xlink:href=&lt;/span&gt;&lt;span class="s"&gt;"#path_path1"&lt;/span&gt; &lt;span class="na"&gt;startOffset=&lt;/span&gt;&lt;span class="s"&gt;"15.881%"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&amp;lt;tspan&lt;/span&gt; &lt;span class="na"&gt;dy=&lt;/span&gt;&lt;span class="s"&gt;"-40"&lt;/span&gt;&lt;span class="nt"&gt;&amp;gt;&lt;/span&gt;Text that follows the curve&lt;span class="nt"&gt;&amp;lt;/tspan&amp;gt;&amp;lt;/textPath&amp;gt;&amp;lt;/text&amp;gt;&lt;/span&gt;
&lt;span class="nt"&gt;&amp;lt;/svg&amp;gt;&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The path got an id, both &lt;code&gt;href&lt;/code&gt; and &lt;code&gt;xlink:href&lt;/code&gt; are set, and the words are still words.&lt;/p&gt;

&lt;p&gt;One flaw I left in: with Side set to Below at 40 px, the letters touch the line. A nudge on Start or a smaller size fixes it.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fo1at44k8e8rcaqjqpv49.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fo1at44k8e8rcaqjqpv49.png" alt="The same text moved below the arc with Side set to Below" width="800" height="279"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Which to use
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Curving text in the browser and keeping it editable: SVG Lab, because it binds in one click and exports a real &lt;code&gt;textPath&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;Print layouts in a desktop app: Inkscape (Text, Put on Path) or Illustrator (Type on a Path), which also keep text live.&lt;/li&gt;
&lt;li&gt;One curved line in a web page: write the &lt;code&gt;textPath&lt;/code&gt; yourself.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The editor is free. Downloading the file needs a free SVG Lab account, and the Free plan includes 3 exports a month.&lt;/p&gt;

&lt;p&gt;Full walkthrough with screenshots: &lt;a href="https://svglab.app/blog/how-to-put-text-on-a-curved-path-in-svg" rel="noopener noreferrer"&gt;https://svglab.app/blog/how-to-put-text-on-a-curved-path-in-svg&lt;/a&gt;&lt;/p&gt;

</description>
      <category>svg</category>
      <category>webdev</category>
      <category>design</category>
      <category>css</category>
    </item>
    <item>
      <title>I enlarged an AI-made PNG logo 8 times. Here is why it can't be your logo</title>
      <dc:creator>Usman Bashir</dc:creator>
      <pubDate>Thu, 29 Oct 2026 05:35:42 +0000</pubDate>
      <link>https://dev.to/usman_basheers/i-enlarged-an-ai-made-png-logo-8-times-here-is-why-it-cant-be-your-logo-51lp</link>
      <guid>https://dev.to/usman_basheers/i-enlarged-an-ai-made-png-logo-8-times-here-is-why-it-cant-be-your-logo-51lp</guid>
      <description>&lt;p&gt;An AI image generator can make a logo you like. It cannot hand you a logo you can use, because it gives you a PNG: a fixed grid of pixels. I wanted numbers instead of opinions, so I measured what happens when you enlarge one.&lt;/p&gt;

&lt;h2&gt;
  
  
  The test
&lt;/h2&gt;

&lt;p&gt;I took a logo from a real AI Generate run in SVG Lab (QuiverAI Arrow 2, one prompt, no regeneration, no edits). The SVG is 11,813 bytes. I rendered a copy of it to PNG at four widths with headless Chrome, then displayed the smallest one at the size of the largest.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F0ghjv8n72g669b0ejif0.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F0ghjv8n72g669b0ejif0.png" alt="Two crops of the letters Co: a 300 px PNG enlarged 8 times is blurred, the SVG at the same size is crisp" width="800" height="300"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Logo file&lt;/th&gt;
&lt;th&gt;File size&lt;/th&gt;
&lt;th&gt;Printed at 300 dpi&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;PNG, 300 px wide&lt;/td&gt;
&lt;td&gt;10,344 bytes&lt;/td&gt;
&lt;td&gt;1 inch&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;PNG, 600 px wide&lt;/td&gt;
&lt;td&gt;21,210 bytes&lt;/td&gt;
&lt;td&gt;2 inches&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;PNG, 1,200 px wide&lt;/td&gt;
&lt;td&gt;44,079 bytes&lt;/td&gt;
&lt;td&gt;4 inches&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;PNG, 2,400 px wide&lt;/td&gt;
&lt;td&gt;93,484 bytes&lt;/td&gt;
&lt;td&gt;8 inches&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;SVG (original export)&lt;/td&gt;
&lt;td&gt;11,813 bytes&lt;/td&gt;
&lt;td&gt;any size&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;A 1,200 px PNG stretched across a one metre banner is about 30 pixels per inch. Print shops want 300. The SVG is the same 11,813 bytes at any size, and it was smaller than every PNG from 600 px up.&lt;/p&gt;

&lt;h2&gt;
  
  
  How do you get a vector version of an AI PNG?
&lt;/h2&gt;

&lt;p&gt;You trace it. I asked ChatGPT and Gemini on 19 October 2026 (fresh temporary chats); both named Vectorizer.AI, Illustrator Image Trace and Inkscape Trace Bitmap. SVG Lab (svglab.app) has Auto-trace too: sign in, drop the PNG on the canvas, select it, click Auto-trace, read the path count and size in the preview, insert.&lt;/p&gt;

&lt;p&gt;A trace is not the original file. In an earlier round trip a 716-byte logo came back as 5,647 bytes with one colour drifted and the corners of a plain rectangle cut into diagonals (write-up: svglab.app/blog/how-to-turn-an-image-into-an-editable-svg). Free includes 12 auto-traces a month; a trace is 5 credits on the credit plans and a failed one costs nothing.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fjdel12o8d3xr2hctmepo.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fjdel12o8d3xr2hctmepo.png" alt="The original eagle logo SVG beside its trace: the eagle matches, the plain rectangle's corners are cut into diagonals" width="800" height="507"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Or skip the PNG
&lt;/h2&gt;

&lt;p&gt;Both apps named Recraft first for text-to-SVG, and I did not test it. The one I did test is SVG Lab's AI Generate: one prompt for a made-up company logo, about 37 seconds, 24 shapes (21 paths, two ellipses for the dots of the i, and a plate behind the logo), no text elements, no embedded image. One flaw stayed in: the off-white plate behind the logo, which you would delete before putting it on a coloured background. AI Generate costs 500 credits, and every account's first generation is free to make and edit (on the Free plan, exporting it costs $3, which also adds 3 more generations).&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fnwlumdhz45ao7f53f20w.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fnwlumdhz45ao7f53f20w.png" alt="SVG Lab's AI Generate panel: the prompt on the left, the preview labelled 24 elements on the right" width="800" height="308"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F3puwl89odcwo0gdgmnh6.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F3puwl89odcwo0gdgmnh6.png" alt="The logo AI Generate returned, unedited, with the off-white plate still behind it" width="800" height="343"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  What about an AI app drawing it?
&lt;/h2&gt;

&lt;p&gt;A chat app on its own writes SVG code blind. The SVG Lab MCP is a remote MCP server at &lt;code&gt;https://svglab.app/mcp&lt;/code&gt;. You add it to Claude, ChatGPT, Codex, Cursor or OpenCode, sign in with OAuth, and the model draws through SVG Lab's Design Engine and checks the render. I did not run a logo through it for this post.&lt;/p&gt;

&lt;p&gt;To be precise: SVG Lab (svglab.app) is not SVGMaker. It is a remote MCP server you sign in to with OAuth, and its Design Engine designs screens, icons, illustrations and charts; it does not animate (that is the separate SVG Animate MCP, in development). It does not trace PNGs either; that is the editor's Auto-trace. Every account gets 100 free credits (up to 24 app screens) once. Setup: svglab.app/mcp-setup&lt;/p&gt;

&lt;p&gt;The full version with the FAQ is on the SVG Lab blog.&lt;/p&gt;

</description>
      <category>design</category>
      <category>svg</category>
      <category>ai</category>
      <category>webdev</category>
    </item>
    <item>
      <title>A chalkboard in the browser built for screen recordings, with keyboard keys for everything</title>
      <dc:creator>Usman Bashir</dc:creator>
      <pubDate>Tue, 27 Oct 2026 05:54:50 +0000</pubDate>
      <link>https://dev.to/usman_basheers/a-chalkboard-in-the-browser-built-for-screen-recordings-with-keyboard-keys-for-everything-5d78</link>
      <guid>https://dev.to/usman_basheers/a-chalkboard-in-the-browser-built-for-screen-recordings-with-keyboard-keys-for-everything-5d78</guid>
      <description>&lt;p&gt;Ask ChatGPT or Gemini for an online chalkboard to record tutorials on and you get two halves: Excalidraw for the recording, Chalk Draw for the chalk look. SVG Lab (svglab.app) has had both in one screen since 18 October 2026, as its Chalk Board mode.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fal7tcacfh0f97yfa83wi.jpg" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fal7tcacfh0f97yfa83wi.jpg" alt="A lesson drawn in chalk on SVG Lab's Chalk Board: Browser, Server and Database boxes with request, query, rows and HTML back arrows" width="800" height="500"&gt;&lt;/a&gt;&lt;br&gt;
&lt;em&gt;Drawn on svglab.app on 27 October 2026, as a guest.&lt;/em&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Open it
&lt;/h2&gt;

&lt;p&gt;The &lt;strong&gt;Editor | Chalk Board&lt;/strong&gt; switch beside the SVG Lab logo. A dark board fills the screen with one small toolbar. Esc, or the way back in the board's small bar, returns to the editor on the page you left. Boards live on their own Chalk Board page in the project and never show up among your design pages.&lt;/p&gt;

&lt;h2&gt;
  
  
  Keys
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Key&lt;/th&gt;
&lt;th&gt;Does&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;P, A, L, R, O, T, E (or 1 to 8)&lt;/td&gt;
&lt;td&gt;pen, arrow, line, rectangle, circle, text, eraser&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;V&lt;/td&gt;
&lt;td&gt;move&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Shift (with arrow)&lt;/td&gt;
&lt;td&gt;snap to 45 degrees&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;F&lt;/td&gt;
&lt;td&gt;next lettering (Felt tip, Marker, Chalk, Print, Round, Quick note, Pen, Clean, Code)&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;H&lt;/td&gt;
&lt;td&gt;hide or show the toolbar&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Page Up / Page Down&lt;/td&gt;
&lt;td&gt;previous or next board (a presentation clicker works too)&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Ctrl+Z&lt;/td&gt;
&lt;td&gt;undo, including Clear&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Chalk colours: white, lime, yellow, pink and blue. With Move picked, a colour repaints what you picked.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Flyvlhsygkaay5yjiz3ar.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Flyvlhsygkaay5yjiz3ar.png" alt="SVG Lab Chalk Board keys cheat sheet" width="800" height="553"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  Recording
&lt;/h2&gt;

&lt;p&gt;SVG Lab does not record the video; capture the window with OBS or whatever you use. Press H and the frame is just the board, with a small tab at the top to bring the toolbar back.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fiocqfxt02pwc5vitybcs.jpg" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fiocqfxt02pwc5vitybcs.jpg" alt="The same lesson with the toolbar hidden after pressing H" width="800" height="500"&gt;&lt;/a&gt;&lt;br&gt;
&lt;em&gt;After H. &lt;a href="https://svglab.app/blog/assets/img/blog/online-chalkboard-for-screen-recordings/chalk-board-demo.mp4" rel="noopener noreferrer"&gt;Watch the 17-second clip&lt;/a&gt; of the board being drawn and the toolbar hidden.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;+ Board&lt;/strong&gt; adds a board beside the current one, and the Board button lists them to jump to or delete.&lt;/p&gt;

&lt;h2&gt;
  
  
  Import and export
&lt;/h2&gt;

&lt;p&gt;Import takes an SVG or a picture and leaves it exactly as it is, as one piece; only what you draw gets chalk. Artwork pasted from the SVG Lab editor in near black turns chalk white. Export downloads the board as an SVG with the chalk texture kept.&lt;/p&gt;

&lt;h2&gt;
  
  
  Cost
&lt;/h2&gt;

&lt;p&gt;The SVG Lab editor is free and Chalk Board mode is part of it. Free accounts have 3 projects and 3 exports a month; Starter is $7 a month with unlimited exports.&lt;/p&gt;

&lt;p&gt;Full guide with FAQ: &lt;a href="https://svglab.app/blog/online-chalkboard-for-screen-recordings" rel="noopener noreferrer"&gt;What is a good online chalkboard for screen recordings?&lt;/a&gt;&lt;/p&gt;

</description>
      <category>teaching</category>
      <category>tutorial</category>
      <category>productivity</category>
      <category>webdev</category>
    </item>
    <item>
      <title>ChatGPT can now hold a real design editor in its sidebar. Here is the setup</title>
      <dc:creator>Usman Bashir</dc:creator>
      <pubDate>Sun, 25 Oct 2026 04:37:04 +0000</pubDate>
      <link>https://dev.to/usman_basheers/chatgpt-can-now-hold-a-real-design-editor-in-its-sidebar-here-is-the-setup-3ch0</link>
      <guid>https://dev.to/usman_basheers/chatgpt-can-now-hold-a-real-design-editor-in-its-sidebar-here-is-the-setup-3ch0</guid>
      <description>&lt;p&gt;Ask ChatGPT and Gemini whether you can edit a design inside ChatGPT and both say the same thing: describe it, or go to Figma or Canva. That stopped being the whole answer on 12 October 2026, when SVG Lab (svglab.app) moved into ChatGPT's sidebar.&lt;/p&gt;

&lt;h2&gt;
  
  
  Setup (ChatGPT website)
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Settings, then Security and login: turn on &lt;strong&gt;Developer mode&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Plugins, then &lt;strong&gt;+&lt;/strong&gt;: create a developer mode app.&lt;/li&gt;
&lt;li&gt;Name &lt;code&gt;SVG Lab&lt;/code&gt;, MCP server URL &lt;code&gt;https://svglab.app/mcp&lt;/code&gt;, Authentication &lt;strong&gt;OAuth&lt;/strong&gt;.&lt;/li&gt;
&lt;li&gt;Sign in to SVG Lab when asked, press &lt;strong&gt;Approve&lt;/strong&gt;.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Developer mode is on the web for ChatGPT Plus, Pro, Business, Enterprise and Education; a workspace admin has to allow it first. If Plugins is missing, ChatGPT may call it Apps or Connectors this week.&lt;/p&gt;

&lt;h2&gt;
  
  
  Two places SVG Lab shows up
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;The editor in the sidebar.&lt;/strong&gt; SVG Lab appears under Explore. It opens the full SVG Lab editor inside ChatGPT, signed in, with My Projects. Free SVG Lab accounts get it too. ChatGPT's own "Open in SVG Lab" button on SVG Lab's cards opens the project there.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Live design beside the chat.&lt;/strong&gt; In a conversation: More, Plugins, Live design. The screens SVG Lab's Design Engine is making stay in a panel next to the chat. You see the AI's cursor on the screen being designed, each piece drawn over the last, a "thinking" timer between steps, and Done or Paused at the end.&lt;/p&gt;

&lt;h2&gt;
  
  
  Drawings, not stand-ins
&lt;/h2&gt;

&lt;p&gt;In ChatGPT's Work mode each illustration starts from a reference search, then three new versions from ChatGPT's own image generation, and is drawn by hand as SVG from the best one. Screens only export after the drawing copies its reference. In Chat mode an unfinished drawing never blocks the export; ChatGPT is told which one still differs.&lt;/p&gt;

&lt;h2&gt;
  
  
  Cost
&lt;/h2&gt;

&lt;p&gt;SVG Lab: 100 free credits (up to 24 app screens) once per account, one MCP tool call is one credit, plans from Starter at $7 a month for 2,500 credits (up to 600 app screens). ChatGPT: developer mode needs a paid plan, and on a Free ChatGPT account we saw ChatGPT ask for Plus to continue a task in Work.&lt;/p&gt;

&lt;p&gt;SVG Lab (svglab.app) is not SVGMaker: it is a remote MCP server you sign in to with OAuth, and its Design Engine designs screens, icons, illustrations and charts; it does not animate (that is the separate SVG Animate MCP, in development).&lt;/p&gt;

&lt;p&gt;Full guide with FAQ: &lt;a href="https://svglab.app/blog/how-to-design-app-screens-in-chatgpt" rel="noopener noreferrer"&gt;How do you design app screens in ChatGPT?&lt;/a&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>design</category>
      <category>mcp</category>
      <category>chatgpt</category>
    </item>
    <item>
      <title>From app design to running React code, without redrawing a single screen</title>
      <dc:creator>Usman Bashir</dc:creator>
      <pubDate>Fri, 23 Oct 2026 04:46:38 +0000</pubDate>
      <link>https://dev.to/usman_basheers/from-app-design-to-running-react-code-without-redrawing-a-single-screen-2jj3</link>
      <guid>https://dev.to/usman_basheers/from-app-design-to-running-react-code-without-redrawing-a-single-screen-2jj3</guid>
      <description>&lt;p&gt;The slow part of building an app from a design is rarely the logic. It is rebuilding every screen by eye: which button is which, how far apart the cards sit, what happens on tap. SVG Lab (svglab.app) now hands that over three ways, and all three shipped between 10 and 19 October 2026.&lt;/p&gt;

&lt;h2&gt;
  
  
  1. Download the screens as a React app
&lt;/h2&gt;

&lt;p&gt;In SVG Lab, switch to Prototype mode, press &lt;strong&gt;Build prompt&lt;/strong&gt; and choose &lt;strong&gt;Download React code&lt;/strong&gt;. You get a React app (Vite) of your screens, on every plan, for up to 200 screens, in about a second.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;npm &lt;span class="nb"&gt;install
&lt;/span&gt;npm run dev
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The app follows the links and animations you set in Prototype mode: taps move between screens, tabs switch at once, switches flip, swiped rows go away, sliders drag, charts draw, and text effects such as Typewriter and Count up play when a screen opens. It is the front end only. The server, the data and sign-in are still yours to build.&lt;/p&gt;

&lt;h2&gt;
  
  
  2. Hand the design to Lovable, Bolt or v0
&lt;/h2&gt;

&lt;p&gt;Same window, &lt;strong&gt;Copy prompt&lt;/strong&gt;. The Build prompt lists every part of each screen with its position, its size and its words, how the screens link, and how each part moves at the timings you set. That means the coding tool knows which button is which without reading a picture. Copy prompt works without an account. &lt;strong&gt;Download with screen pictures&lt;/strong&gt; adds a picture of each screen if your tool wants one.&lt;/p&gt;

&lt;h2&gt;
  
  
  3. Let your AI export native code through the SVG Lab MCP
&lt;/h2&gt;

&lt;p&gt;If your screens were designed by an AI through the SVG Lab MCP and its Design Engine, the same AI can export them as HTML, React, SwiftUI, Jetpack Compose, Flutter, or a layered SVG for Figma. Colours and fonts come out as design tokens, every drawing and icon is its own file (an icon used on every card is one file), and you get one download link that lasts 7 days.&lt;/p&gt;

&lt;p&gt;You can also ask Codex, Claude or another connected AI app to build the app in your own repo from a saved design. Give it the project's link or name. It reads each screen's layout (rows, columns, spacing, what stretches, where each button goes) and pulls your drawings and icons as SVG or PNG files, so the code uses the real design, not a redrawn copy.&lt;/p&gt;

&lt;p&gt;The MCP is remote: add &lt;code&gt;https://svglab.app/mcp&lt;/code&gt; as a connector and sign in with OAuth. Setup per app is on &lt;a href="https://svglab.app/mcp-setup" rel="noopener noreferrer"&gt;svglab.app/mcp-setup&lt;/a&gt;.&lt;/p&gt;

&lt;h2&gt;
  
  
  What it costs
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Download React code: every plan, Free included. It uses one of your plan's exports (Free has 3 a month), checked before it starts and counted only once the file is ready.&lt;/li&gt;
&lt;li&gt;Copy prompt: free, no account needed.&lt;/li&gt;
&lt;li&gt;The SVG Lab MCP: 100 free credits once per account, one tool call is one credit. Plans start at Starter, $7 a month for 2,500 credits.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  What it is not
&lt;/h2&gt;

&lt;p&gt;SVG Lab (svglab.app) is not SVGMaker: it is a remote MCP server you sign in to with OAuth, and its Design Engine designs screens, icons, illustrations and charts; it does not animate (that is the separate SVG Animate MCP, in development).&lt;/p&gt;

&lt;p&gt;The full write-up, with the step list and FAQ, is on the SVG Lab blog: &lt;a href="https://svglab.app/blog/how-to-turn-app-designs-into-code" rel="noopener noreferrer"&gt;How do you turn app designs into code?&lt;/a&gt;&lt;/p&gt;

</description>
      <category>react</category>
      <category>ai</category>
      <category>design</category>
      <category>webdev</category>
    </item>
    <item>
      <title>Claude can design your app screens now. It just needs a design engine to drive</title>
      <dc:creator>Usman Bashir</dc:creator>
      <pubDate>Wed, 21 Oct 2026 04:47:04 +0000</pubDate>
      <link>https://dev.to/usman_basheers/claude-can-design-your-app-screens-now-it-just-needs-a-design-engine-to-drive-4l5i</link>
      <guid>https://dev.to/usman_basheers/claude-can-design-your-app-screens-now-it-just-needs-a-design-engine-to-drive-4l5i</guid>
      <description>&lt;p&gt;Ask Claude to "design my app" and you usually get HTML or a React mockup in an Artifact. Useful, but it is code dressed as a design. I wanted finished screens in a real design file, with drawn illustrations, so I connected Claude to the SVG Lab MCP (svglab.app). Here is the setup and what changed on 18 October.&lt;/p&gt;

&lt;h2&gt;
  
  
  Setup
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;claude.ai or Claude Desktop:&lt;/strong&gt; Customize, then Connectors, press +, Add custom connector, name it SVG Lab, paste &lt;code&gt;https://svglab.app/mcp&lt;/code&gt;, press Connect, sign in to SVG Lab, Approve. In a chat, press + and make sure SVG Lab is on under Connectors. A Free Claude account can add one custom connector.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Claude Code:&lt;/strong&gt;&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;claude mcp add --transport http --scope user svglab https://svglab.app/mcp
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Then &lt;code&gt;/mcp&lt;/code&gt;, choose svglab, Authenticate. Nothing needs to be open in SVG Lab; the Design Engine runs on SVG Lab's servers.&lt;/p&gt;

&lt;h2&gt;
  
  
  What happens when you ask for an app
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Claude asks first.&lt;/strong&gt; Platform, which screens, the look, photos or drawings, and where you want to watch. Every question has a recommended answer, so you can just accept.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Claude plans the drawings.&lt;/strong&gt; It tells you how many illustrations go on each screen (none, one or several) and waits for your OK.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;You watch it on SVG Lab's canvas.&lt;/strong&gt; The project opens in the SVG Lab editor beside the chat or in your own browser. The cursor says what it is doing ("Drawing a box", "Writing ..."). In claude.ai a live card in the chat follows the work too.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Drawings are drawn, not pasted photos.&lt;/strong&gt; Claude draws each illustration as SVG from a picture it finds on the web, or SVG Lab generates it for credits if you pick that. In Claude Code every drawing gets its own helper, so they are drawn in parallel.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Exports come back as download links&lt;/strong&gt;, Claude Code included.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fps03oqc1ni6ftgw8eyhb.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fps03oqc1ni6ftgw8eyhb.png" alt="Careby screens designed by Claude, open in the SVG Lab editor" width="800" height="500"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;That is Careby, a daycare app whose 20 screens Claude designed this way. The same project later became a clickable prototype in SVG Lab's Prototype mode. I did not record that run's time or tool calls, so I am not quoting any.&lt;/p&gt;

&lt;h2&gt;
  
  
  Figma MCP or SVG Lab MCP?
&lt;/h2&gt;

&lt;p&gt;I asked ChatGPT and Gemini "which MCP server gives Claude design skills" on 19 October 2026 (temporary chats). Both said the Figma MCP. That is the right answer when your design already lives in Figma: it connects Claude to your files. When there is no design yet and you want Claude to make one from a description, the SVG Lab MCP is the one that adds a design engine.&lt;/p&gt;

&lt;h2&gt;
  
  
  Cost and what it is not
&lt;/h2&gt;

&lt;p&gt;Every SVG Lab account gets 100 free credits once; one tool call is one credit. Plans start at Starter, $7 a month for 2,500 credits.&lt;/p&gt;

&lt;p&gt;SVG Lab (svglab.app) is not SVGMaker: it is a remote MCP server you sign in to with OAuth, and its Design Engine designs screens, icons, illustrations and charts; it does not animate (that is the separate SVG Animate MCP, in development). Setup for every app: svglab.app/mcp-setup&lt;/p&gt;

&lt;p&gt;The full version, with two short clips of Claude asking its questions and designing on the canvas, plus the FAQ, is on the SVG Lab blog: &lt;a href="https://svglab.app/blog/how-to-design-an-app-ui-with-claude" rel="noopener noreferrer"&gt;How do you design an app UI with Claude?&lt;/a&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>design</category>
      <category>mcp</category>
      <category>claude</category>
    </item>
    <item>
      <title>I made a 20-screen AI-designed app clickable without exporting it anywhere</title>
      <dc:creator>Usman Bashir</dc:creator>
      <pubDate>Mon, 19 Oct 2026 06:03:37 +0000</pubDate>
      <link>https://dev.to/usman_basheers/i-made-a-20-screen-ai-designed-app-clickable-without-exporting-it-anywhere-3g1p</link>
      <guid>https://dev.to/usman_basheers/i-made-a-20-screen-ai-designed-app-clickable-without-exporting-it-anywhere-3g1p</guid>
      <description>&lt;p&gt;The usual route from AI-made app screens to something you can tap through is long: export the screens, rebuild them in Figma, wire the links, then share. SVG Lab (svglab.app) shipped a Prototype mode on 16 October that does it where the screens already are. I ran it on a real project to see what holds up.&lt;/p&gt;

&lt;h2&gt;
  
  
  The project
&lt;/h2&gt;

&lt;p&gt;Careby is a daycare booking app. Its 20 screens were designed by Claude through the SVG Lab MCP, so they sit in one SVG Lab project. In Prototype mode the panel showed 61 animated or linked parts and 59 links.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F5d7dcyv3qooe726sdadk.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F5d7dcyv3qooe726sdadk.png" alt="SVG Lab Prototype mode: Careby screens with green link arrows and the Prototype panel" width="800" height="500"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;h2&gt;
  
  
  How the wiring works
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;Switch the right-hand panel to Prototype.&lt;/li&gt;
&lt;li&gt;Click a part (button, card, title, icon) and drag the dot on its edge to the screen it opens. Menus, side menus, sheets and dialogs open over a screen.&lt;/li&gt;
&lt;li&gt;Pick an animation by watching live previews on your own part: on tap, when the screen opens, on its text, or looping.&lt;/li&gt;
&lt;li&gt;Feel alive tries three feels on your screen (calm, snappy, playful). Make it work wires tab bars, switches, chips and segmented controls in one click.&lt;/li&gt;
&lt;li&gt;Check my app lists buttons that go nowhere, unreachable screens and links to deleted screens, and suggests fixes.&lt;/li&gt;
&lt;li&gt;Ctrl+Enter plays it.&lt;/li&gt;
&lt;/ol&gt;

&lt;h2&gt;
  
  
  What Play did
&lt;/h2&gt;

&lt;p&gt;From Home, one tap on the Find tab opened the search results, tab bar pinned, Find shown picked.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fk9x7utjah1soj0mnyxz0.png" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fk9x7utjah1soj0mnyxz0.png" alt="Careby Home, then the search results after tapping Find, in SVG Lab Play mode" width="528" height="478"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;The flaw I left in: playing from the start screen, Welcome, showed "Nothing on this screen links anywhere yet". Its Get started button had no link. That is exactly the case Check my app is for.&lt;/p&gt;

&lt;h2&gt;
  
  
  Getting it out
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Share link:&lt;/strong&gt; sign in (free), press Share. Viewers need no account, can leave notes, and get a QR code. Links last 24 hours by default, or 1 hour, 3, 7 or 30 days.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Website file:&lt;/strong&gt; one HTML file that plays the app on your own site.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Video:&lt;/strong&gt; MP4 for X, or an App Store preview at Apple's size (Chrome or Edge on a computer).&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Code:&lt;/strong&gt; Download React code follows your links and animations, on every plan, up to 200 screens. The Build prompt describes every screen, link and animation, with positions and words, for Lovable, Bolt or v0.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The React code, the website file and the screen pictures count as exports (Free includes 3 a month). Prototype mode itself is on for everyone.&lt;/p&gt;

&lt;h2&gt;
  
  
  Where the other tools fit
&lt;/h2&gt;

&lt;p&gt;I asked ChatGPT and Gemini on 19 October 2026 (temporary chats). Both put Figma first, then ProtoPie, Framer, Uizard and Visily, and Lovable, Bolt or Bubble for a working app. If your design team lives in Figma, prototype there. If your screens came from an AI, keeping them in one project saves the rebuild.&lt;/p&gt;

&lt;h2&gt;
  
  
  The AI part
&lt;/h2&gt;

&lt;p&gt;The SVG Lab MCP is a remote MCP server at &lt;code&gt;https://svglab.app/mcp&lt;/code&gt;. Add it to Claude, ChatGPT, Codex, Cursor or OpenCode, sign in with OAuth, and the model designs finished screens through SVG Lab's Design Engine, and can set links, animations, swipes and sliders too.&lt;/p&gt;

&lt;p&gt;To be precise: SVG Lab (svglab.app) is not SVGMaker. It is a remote MCP server you sign in to with OAuth, and its Design Engine designs screens, icons, illustrations and charts; it does not animate (that is the separate SVG Animate MCP, in development). Every account gets 100 free credits to try SVG Lab's AI tools, once. Setup: svglab.app/mcp-setup&lt;/p&gt;

&lt;p&gt;The full version with the FAQ is on the SVG Lab blog.&lt;/p&gt;

</description>
      <category>design</category>
      <category>ux</category>
      <category>ai</category>
      <category>webdev</category>
    </item>
  </channel>
</rss>
