Icecat Studio Sprint 104 Release Notes: Verified Publishing and Full Accessibility Coverage

Sprint 104 is about making what we already have more dependable. Publication status is now verified rather than assumed; our accessibility program now reaches the last pages of the studio interface; and both editors now use a shared set of design tokens. Behind the scenes, we continued work on the AI Assistant, the MCP Server, and the new Story Builder canvas.

Publication Status Check

Publishing to the Icecat Network could previously return a successful response while the story did not actually appear — or was removed shortly afterward by validation on the receiving side. The feedback status alone was not sufficient evidence.

Publication is now verified directly. After publishing, the platform requests the story back and confirms that the content is genuinely being served. “Published” in Icecat Studio now means checked on the network, not merely sent without error.

Accessibility: full interface coverage

Earlier sprints brought the asset pages and the story and sandbox editors up to WCAG 2.2 AA and produced a set of internal guidelines from that work. In Sprint 104, we applied the same standard to every remaining page of the studio interface — brands, categories, companies, components, countries, locales, layouts, partners, products, fonts, moderation, reports, history, users and groups, and the login flow.

Each area received the same treatment: proper heading structure, accessible names for icon-only controls, spoken announcements for asynchronous errors and status changes, contrast floors on text and icons, and single-column reflow at small viewport widths.

Three structural improvements go beyond the baseline:

• Keyboard alternatives to drag-only reordering. Brand publish priority, and the component grid within a group can now be reordered entirely from the keyboard, through the same save path as dragging.
• Route unification. Component groups have moved under the components section, with redirects from the previous URLs.
• A shared info-button component replaces a collection of individually styled status tooltips.

All of it was verified live, with real keyboard input, in both light and dark themes.

Design system tokens in Sandbox and Story Builder

Both editors carried a long history of hardcoded colors — this button one shade, that panel another, a third almost identical but not quite. Color is now resolved through design tokens with semantic roles — panel surface, border, icon, error state — rather than being set on individual elements.

The practical result is twofold: the two editors finally look consistent, and the accessibility work above will not drift over time, because contrast floors live in the tokens rather than in each component.

Icecat Studio AI Assistant & MCP Server

• AI Assistant: The assistant entered internal testing — a chat panel inside the studio that takes instructions in plain language: finding assets, creating stories from a URL, file, or pasted text, editing modules, and publishing. It is aware of the page you are on, confirms every change before applying it, and works within your own permissions.
• MCP Server: Development continued on the server that will expose Icecat Studio capabilities to external AI agents via the Model Context Protocol — reusing the AI Assistant’s tool layer, with API-key authorization scoped to each user’s existing permissions and limits.

Story Builder: Central Section – In progress

A new canvas interaction layer: Work continued on moving the central area of the Story Builder — the part where you see the story itself — from an embedded frame into the application proper, so the editor understands exactly what you are looking at and what you click. Modules and individual elements can be highlighted and selected directly on the canvas; a double-click opens the element’s editor, and the module tree expands to the same level of detail and stays synchronized with the canvas in both directions.