Changelog

What shipped, and why. Newest first.

31 October 2026

Who AI search recommends instead of you

After the free scan, ask OpenAI's GPT-5.6, with web search on, one real buying question for your main category and see which brands it names, in about 30 seconds and without leaving an email. Store reports show the same answer. The question uses a specific product type, so "Womens" or "Accessories" never become the category.

Emailing a report works again

A security check behind our proxy rejected every request to email a scan report. Fixed.

Answer checks on GPT-5.6

In-app answer checks moved from GPT-4o mini to GPT-5.6 Luna, which runs several web searches per answer like ChatGPT search does. Questions run in parallel, so a check takes about as long as before.

Free fixes 20 products end to end

Free now fixes category, vendor and product type on up to 20 products, with undo. Growth fixes the whole catalog and adds AI-written descriptions.

The first scan finishes even if you leave

The first scan now runs on our server when you open the app, so the result (and its email) is there when you come back.

30 October 2026

AI shopping channel readiness

Shopify now sends catalogs to ChatGPT, Copilot and Perplexity, and Google shows products in AI Mode. The scan shows, per channel, how many products can appear, how many are excluded and why, with a link to the fix.

Shopify categories in bulk

Products without a Standard Product Taxonomy category get one suggested from Shopify's own category list. Preview, apply, undo. Free assigns 10; Growth covers the whole catalog. The scan also flags categorized products whose attributes (color, material, size) are all empty.

Why the AI picked a competitor

When an answer check names someone else, Endcap reads the product page the answer cited and lists what it publishes that your products lack, each linked to its fix.

First scan runs on install

Opening the app for the first time starts the catalog scan instead of waiting for a click.

2 August 2026

Bulk fixes

The app can now write vendor, product type and descriptions across the catalog. Every change is previewed before it is applied and any batch reverts in one click, because an app with write access to a live catalog earns trust by being reversible.

Descriptions written from product data

Drafts are generated strictly from each product's own fields. Internal tags, types and variant option values are filtered out first, after testing produced customer-facing copy containing warehouse codes.

Publication detection made robust

Products on password-protected and development stores were being reported as unpublished, so publication now takes two independent signals before a product is called blocked.

1 August 2026

AI visibility tracking

Buying questions are derived from your catalog, put to a live assistant, and reported as unbranded share of voice with a competitor leaderboard. The question set is frozen between runs so the trend is comparable.

Housekeeping detection

Content blocks, test rows and zero-priced helper products from bundle apps are now classified and excluded from your score, after early scans reported an entire content system as broken products.

First release

27 deterministic checks against Shopify's published Catalog requirements, 22 per product and 5 across the store, each finding linking to the document it came from.