Imejis.io

The built-in assistant takes a plain description and produces a layout, headline, subline and price field. It picks a structure that fits the content, then lets you refine with follow-up prompts. The same assistant is reachable through the agent connection, so a request typed in Claude can create a design without opening the editor.

A single design can hold as many sizes as you need, chosen from presets or set to any custom width and height. Each new size starts as an auto-reflow of the master, which keeps margins and relative proportions. Every size shares one configuration and one set of keys, so changing the headline on the master updates the story and LinkedIn versions too.

Every template exposes a render endpoint under render.imejis.io, and every field is a JSON key. A POST request returns the image bytes, a hosted URL or a signed URL in the same response, with no webhook or polling step. Output goes up to 4096 by 4096 pixels, and JPEG quality is adjustable from 1 to 100.

get_render_url returns a permanent signed link that re-renders on each request with new field overrides. That suits email footers, dashboards and any place where the same slot needs fresh numbers. Each call counts as a render, which matters when a link sits on a high-traffic page.

Designs can contain text, images, shapes, tables, QR codes, barcodes, ratings, progress bars and 19 chart types such as bar, line, pie, radar, sankey, treemap, funnel and candlestick. The editor, the agent and the API read from the same component registry, so a preview matches the final render. Agents can query the live catalog before building, which avoids guessed property names.

You can upload typefaces, logos and photos once and reuse them across designs. Uploaded fonts appear in the icon picker for every size and every template in the workspace. That keeps a campaign consistent when several people or agents produce images from the same library.

Edits save automatically as a private draft, and API calls or share links keep serving the published version until you publish. Each publish creates a named version with a preview image, and restoring an older one creates a new version instead of rewriting history.

The same render endpoint works from no-code tools. A new spreadsheet row, form entry or CRM lead can trigger a render and pass the returned image to email, Slack or cloud storage. This covers teams that want automated social media images without writing code or connecting an agent.