Live2026

OG Image Maker

Make a share image or blog cover in a simple view or with full control. Take the colours from your site address or a picture, and preview it on desktop and phone.

Next.jsTypeScriptSVGCanvas

Most share images come from a design tool or a template that decides everything, and the full-control tools bury what most people need. This one opens in a simple view: pick what the image is for (computer or phone), type your words, choose a look and download, with a preview as a link card on a computer and in a chat on a phone and a one-minute guided tour. Switch to Everything for per-text fonts, sizes and colours, backgrounds, patterns, shapes, a frame and a logo. Give it your site address or a picture and it reads the main colours and builds a matching theme with readable text. The image is drawn from your device fonts, so the preview is exactly the file you download. Pictures you choose never leave your browser; a web address you type is sent to the server once to fetch its picture, and nothing is stored.

Case study

Problem

A page goes out with a plain link because making a share image means opening a design tool, or it uses a template that cannot be changed. When a hero image exists, matching its colours by hand is slow and the text ends up unreadable on it.

Approach

Measure real text on a canvas, wrap and fit each text to its box, and draw everything as one SVG that also becomes the PNG or JPG. Reduce a picture to a small thumbnail, find its dominant colours, and build a theme from them with a contrast check on the text.

Key decisions

  • System fonts only. A picture made in the browser cannot load a web font, so using web fonts would make the preview and the file differ.
  • The colour extraction copies colours, not the design of the picture. Fonts and layout stay the visitor's to choose.
  • Contrast is a guarantee, not a hint: when neither white nor near-black text reaches 4.5:1, the background is moved until it does.
  • Break long words between graphemes, not code points, so a Devanagari letter is never separated from its vowel sign.
  • A picture is only ever embedded as a data URL of an image type, so a typed value cannot point the image at another site.
  • Fetching another site picture needs a server, because browsers block it. The route refuses private and internal addresses at connect time, re-checks every redirect, caps size and time, and answers only same-site requests.

On measurable results

No usage data yet. This is a self-initiated tool, not a deployed one.

Highlights

  • Simple view by default, every control behind one Everything switch, plus a guided tour
  • Fully client-side: the title, pictures and image never leave the browser
  • Five texts (title, subtitle, tag line, badge, footer), each with its own font, weight, size, colour, opacity, spacing, capitals and italics
  • Style from a picture: median-cut palette with k-means refinement, a theme with text that is always readable (4.5:1), and a darkening layer sized to the picture
  • Backgrounds, eight patterns, marks, corner shapes, frame, text panel, logo, split layout with a picture panel
  • PNG, JPG or SVG at four sizes or a custom size, with double resolution
  • Long words break between whole characters, so Hindi letters keep their marks
  • Style from a web address: a locked-down server fetch (public hosts only, redirects re-checked, size and time limits, rate limited) reads the page og:image
  • Ready-made starts for computers and phones, and a preview as a desktop link card and a phone chat

Open now · reply within 24 hours

Need something like this built?

I take on freelance and contract frontend work, remote, for clients worldwide.

Prefer email? hello@devstash.me

What happens next

  1. 1

    You write

    A few lines on the goal, the deadline and where the traffic comes from.

  2. 2

    I reply within 24 hours

    With questions, or a quote if the scope is already clear.

  3. 3

    We agree the scope

    A written scope and a fixed quote before any work starts.