wtf is tooka?

tooka is a framework for building social media card carousels, specifically the 1080 by 1350 pixel format used for Instagram style posts. Instead of using a design tool or an image generation model, you describe the cards you want in plain language and a coding AI agent, such as Claude Code, Codex, or Cursor, writes React components that tooka turns into finished cards. Because the text is rendered from real code rather than drawn by an image model, fonts and characters, including non Latin scripts, come out sharp and correctly formed. The project organizes visual identity through brand kits stored in a themes folder, so a set of colors, fonts, and recurring layout pieces can be reused across many carousels, keeping a whole series looking consistent. It also includes a click to comment inspector built into the browser preview: a user can click any element on a card, type a short note like make this red, and then run a command that has the agent apply every pending comment at once and clear the markers. Beyond building cards, tooka can publish them. It connects to Facebook Pages, Instagram, and Threads through Meta's Graph API, and because Meta requires publicly reachable image URLs, tooka uploads the exported PNG images through the Imgbb API so a carousel can be posted without the user hosting images themselves. The tool also drafts a caption for each platform based on the card content, and it tracks token validation and connection status for each linked account. Other included features are an assets manager for images, videos, and fonts, a logo search built on the svgl catalogue, a fullscreen presentation mode with speaker notes and a timer, and a folder based system for organizing many carousels. The finished output is a plain static site that can be deployed to services like Vercel, Cloudflare Pages, Netlify, or Zeabur with no server required. tooka started as a reworked fork of an existing slide deck framework called open slide, keeping its agent driven authoring approach but changing the canvas size, the export format, and adding the entire publishing pipeline to Meta platforms.