Icecat Release Notes 241: Boosting Data Transparency, AI Reliability, and Platform Efficiency

In release 241, we delivered a set of enhancements across Icecat taxonomy, AI integrations, and platform operations. This release focuses on improving transparency, reliability, and efficiency, helping our partners access complete and accurate product information, enabling AI agents to provide smarter answers, and supporting our internal teams with faster, more reliable workflows. For additional details, please refer to the previous Icecat Release Notes.

Improving Transparency and Consistency in Icecat Taxonomy

In this sprint, we focused on strengthening the Icecat taxonomy foundation by improving how feature values are represented, identified, and shared with our partners. These changes are aimed at increasing data transparency, integration reliability, and future extensibility of our taxonomy exports.

Below are the key improvements delivered in this sprint.

Full Transparency of Feature Values and Translations in FeatureValuesVocabularyList.xml

Previously, the FeatureValuesVocabularyList.xml file exposed only a subset of feature values and translations from the Icecat Feature Values Vocabulary. The export logic applied additional filtering rules:

• Feature values without explicit translations (only a key value) were not exported.
• If a locale translation was identical to the key value, that translation was omitted.

As a result, the file contained only “available” or “distinct” translations, not the full picture.

While this approach reduced redundancy, it also meant that partners did not have a complete and transparent view of all feature values and their translations stored in Icecat.

We deliberately changed the philosophy of this reference file.

Starting from this sprint:

• All feature values present in the Feature Values Vocabulary are exported.
• All translations are explicitly declared for every locale: even if a translation is identical to the key value and even if it does not add linguistic variation.
• No feature values are silently filtered out.

In other words, FeatureValuesVocabularyList.xml is now a fully transparent and exhaustive source of truth.

This change ensures that Icecat delivers the most accurate and comprehensive representation of its feature values vocabulary.

Introducing IDs for Restricted (Predefined) Feature Values

Icecat taxonomy already uses IDs for most structural elements:

• Categories
• Features
• Category–Features relations

However, feature values themselves historically did not have IDs. This was especially limiting for restricted (predefined) feature values, which are commonly used in structured product descriptions.

In this sprint, we made the first foundational step toward solving this gap. All restricted feature values now have unique IDs stored in the database. 

This applies to feature types where values are predefined:

• Dropdown
• Multidropdown
• Yes/No
• Yes/No/Optional

At this stage, the IDs are available internally and prepared for future exposure.

Having stable IDs for predefined feature values unlocks important future capabilities:

• Reliable import and export of feature values without ambiguity
• Easier mapping and synchronization across systems
• Better data integrity when values are reused, localized, or updated

This sprint represents an initial but critical step. After further analysis and preparation, we plan to expose these feature value IDs in export files, enable partners to use IDs as stable references instead of text values, and further improve the experience of working with Icecat taxonomy at scale.

These improvements strengthen Icecat’s taxonomy as a robust, transparent, and integration-ready platform, and prepare the ground for even more powerful capabilities in upcoming releases.

Enhancing AI Interactions via the Icecat MCP Server

In this sprint, we delivered several improvements to the Icecat Model Context Protocol (MCP) server, aimed at enabling AI agents to consume data accurately and ensuring that end users interacting through these AI agents receive trustworthy and consistent answers.

Even small refinements to input and output structures can significantly improve AI reasoning and response quality, and this sprint focused on exactly that.

Clearer Locale Handling in AI Outputs

Icecat product content uses a backup logic: if a product asset (manual, safety document, etc.) is not available in the requested locale, it is provided in a backup locale.

Previously, MCP outputs listed a detail about asset locale, which could be misinterpreted by AI agents without additional context.

In this release, we removed raw locale data from outputs until supporting vocabularies are in place. This ensures AI agents only see clear, unambiguous information, reducing the risk of misinterpretation.

Improved Input Naming for AI Tool Calls

All MCP tool input properties are now consistently formatted using snake_case, providing a uniform and predictable interface for AI agents.

Enhanced Image Data Structure

The structure of image outputs in MCP tools has been refined for clarity, making it easier for AI agents to interpret images correctly.

This change improves semantic clarity without affecting the data itself, ensuring AI models can reliably understand and describe images.

These improvements make MCP a more reliable interface for AI agents, ensuring that responses are accurate, predictable, and grounded in real data, users interacting with AI agents receive trustworthy and clear information, and MCP continues to provide a strong foundation for future AI-first features

Other Key Developments In Sprint 241

In sprint 241, we delivered other updates aimed at improving efficiency, reliability, and the overall experience for both Icecat’s internal teams and external partners. These changes make our platform faster, more reliable, and easier to use, while ensuring uninterrupted access to Icecat content.

Helping Editors Work Smarter With Bulk Image Management

Editors can now manage multiple product images at once, a feature previously available only to senior editors.

Benefit: Faster content updates and higher productivity for our editorial team, ensuring products are described and illustrated more efficiently.

Ensuring Reliable System Operations

• Stable Task Processing

We fixed issues that occasionally caused tasks in our system to get “stuck” or behave unpredictably.

Benefit: Internal operations now run smoothly, reducing delays in content updates and system processes that support our partners.

• Continuous Access to Product Images and Multimedia

We introduced an automatic system that ensures images and multimedia remain available even during storage or system interruptions.

Benefit: Channel partners and users can always access product content without disruption.

• Future-Ready Infrastructure

Updated our underlying technology to make future upgrades smoother and more reliable.

Benefit: A stable and maintainable platform that reduces the risk of downtime for users and partners.

Strengthening Security

Upgraded security for user accounts and access to Icecat content.

Benefit: Partners and users can trust that their data and access are safe and secure.

Optimizing Platform Efficiency

• Faster Video Processing

Reduced delays in our video content pipeline.

Benefit: Videos are available more quickly and reliably to partners and users.

• Storage and Resource Management

Cleaned up unused FTP accounts, freeing around 100 GB of storage. Paused unnecessary automated scripts that were consuming resources.

Benefit: More efficient use of system resources, helping the platform run smoother for everyone.

Summary

This release strengthens Icecat’s platform by making product data more transparent and reliable, improving AI-powered interactions, and enhancing overall efficiency and stability. Partners and internal teams will benefit from clearer, more accurate product information, faster editorial workflows, and a platform that is safer, more resilient, and easier to use. Together, these updates pave the way for a smarter, more intuitive Icecat experience.