Icecat Studio Sprint 105 Release Notes: Opening Icecat Studio to the Agent Ecosystem

Sprint 105 shipped one of our anticipated milestones yet: the Studio MCP server went live in production, letting external AI agents operate the platform through a standard protocol. Alongside it, Story-in-Story moved closer to release, the native Story Builder continued reaching full feature parity with the legacy editor, and the AI Assistant’s client panel got a round of usability upgrades based on team feedback and is now close to release.

MCP Server – Live in Production

Following successful testing on dev, the Icecat Studio MCP server is now live. External AI agents and assistants – including Claude itself – can operate Icecat Studio through the standard Model Context Protocol, using the exact same capabilities as our built-in AI Assistant:

• Find and inspect content: Search assets by brand, MPN, or name, and check their stories and publication status across channels.
• Create and author stories: Generate a complete story from a product URL, a data file (JSON/XML/CSV), pasted text, or images – and create new assets from an MPN or ASIN.
• Edit existing content: Update specific details and modules and layout on a story already in progress.
• Manage publication: Move a story to approved and run publication, including an explanation of where – or why not – something got published.
• Answer how-to questions: Pull answers directly from the official Icecat Studio documentation.

Every action still respects the user’s own permissions (brands, languages, categories), and every data-changing action can be gated the same way it is in the app. Also, destructive actions are limited, so you can’t remove stories in bulk or affect the catalog in any other destructive ways.

Setup is easy and user-friendly, takes only a few steps and an API key.

If you’d like to connect your own AI agent to Icecat Studio, reach out to us and we’ll send over the setup instructions.

Amazon – Media API Setup

Amazon’s new Media API had been opened, so we began integrating the piece that unlocks video content for A+ Premium stories:

Tests: We ran a series of submissions against test products in Amazon’s environment to work out the integration details end-to-end – request format, authentication, response handling – rather than relying on documentation alone.
Validation rules: we started encoding Amazon’s own constraints directly into our platform – file size limits, format requirements, and the other technical rules a video module has to satisfy before Amazon will accept it.

Story Builder Central Section – Native Editor Reaching Legacy Parity

The native Story Builder is nearing the closing of its remaining gaps with the legacy iframe-based editor in the development environment:

• Hotspots: Add, drag-reposition, and a full keyboard equivalent (arrow keys to move, Enter to save, Escape to cancel) for hotspots on images, including nested ones.
• Table configuration: A per-cell panel for header/metric rows, highlighted columns, and inserting or deleting rows and columns, with limits enforced and empty tables sized via a grid picker.
• Restricted values: Comparison-table “✔ / ✘” cells are now set by picking a value instead of free text, consistent across the canvas, the tree, and the module editor.
• AI Assist everywhere: Available at the entity, module, and full-story level, with live canvas updates and a persistent processing indicator for story-level runs.
• Full keyboard operability: Action bars behave as proper toolbars (arrow-key navigation, Escape to step up a tier), meeting the platform’s accessibility standards.

With this phase, the editor is almost on par with the legacy page for everyday authoring – the remaining work is drag-and-drop reordering and a production rollout decision.

Icecat Studio AI Assistant – Client Panel Usability Pass

Based on internal feedback, the assistant’s chat panel received a round of usability improvements:

• Expandable workspace: The chat window is no longer size-constrained – it can now open full-page, and the left sidebar can be resized larger.
• Reworked prompt input: Bigger, with auto-resize as you type.
• Clearer suggestions: New-chat cards are now split into Actions and How to categories. In an ongoing chat, a new “What can I do here?” button opens an interactive suggestions panel with the same split.
• Context awareness indicator: A hover hint now shows which part of the platform the assistant is “looking at” – what it can do here, plus a few things possible on other pages.