Icecat Studio – Sprint 101 Release Notes

Sprint 101 was a consolidation sprint. Where Sprint 100 opened several fronts, this release brings them to a working state: the Custom Index File is ready for partner delivery, the Amazon A+ Premium API successfully published test content through the beta, and the AI Assistant moved from creating stories to editing them.

Custom Index File

We completed the on-request index file generation feature. Users of our endpoints can now request an XML or JSON index of all stories grouped by assets, filtered by parameters like brand, category, period, etc.

Generation runs asynchronously: one endpoint initiates the request, a second tracks its status, and a third downloads the resulting file. This keeps large catalogues from blocking the caller and makes the feature usable for full-catalogue exports, not only narrow selections.

The feature originated in a partner request to consume an inventory of available rich content rather than individual stories, and it complements our existing Story XML/JSON API and HTML Package delivery options.

Icecat Studio AI Assistant – Editing and Context Awareness

The AI Assistant received the largest development block of the sprint. The headline change: it no longer only creates stories from scratch — it can now modify existing ones.

Story Mapper v3. Story generation was rewritten from a single large prompt into a per-module pipeline. Users see the generation stages as they progress, and token and cache usage is now tracked per run. This makes output quality more predictable and long stories considerably more reliable.

Component Editing. New assistant tools allow adding and updating individual story components. The design is deliberately conservative: one mutating action at a time, with code-level confirmation before anything is written. The feature is currently in testing, where we are tuning prompts against real editing scenarios and edge cases.

Story Creation Improvements. Layout selection is now automatic, media and text slots are distributed more sensibly across the chosen layout, and error messages have been rewritten in plain language instead of technical output.

Page Context. The assistant is now aware of the page the user is on. It proposes relevant actions for that context and links to the specific documentation section rather than to the documentation root.

Client Panel. Suggestion chips, visible tool-call statuses, panel resizing, message attachments, and a debug view for super users.

Variant Versioning – Unique Naming per Source

We completed the naming logic introduced with campaign variants in Sprint 100. Story variants originating from different sources (Story Vuilder and HTML Package) must now carry unique names without colliding, which removes the ambiguity that appeared when the same variant type was produced by more than one source.

Sandbox Improvements

SVG Import – Gradients: Gradient support for SVG import is implemented for testing. The work follows brand partner feedback on Figma-exported modules, where gradients previously failed to import. Testing on the development environment continues in the next sprint.
Rendering Issues: A set of fixes addressing differences between editing mode and final rendered output. Work in progress.
Canva Import (SVG & PDF): Development continues on importing Canva exports directly into the Sandbox, reducing the manual preparation step on the brand side.