Changelog
Every notable change since v9 with the date and the highlight. Versions ship roughly weekly. The latest is v19; older versions remain documented for users on slower release cadences and for context on when a feature shipped.
Versions ship roughly weekly. The latest is v19. Highlights below.
Pingback protocol GA. API early access broader. Personas regenerate flow polished.
- Pingback protocol moves from preview to GA. Cloudflare Worker, Vercel Edge Middleware, Fastly VCL, and Lambda@Edge templates published.
- API early access expanded to 100 keys. Webhooks for audit.completed, prompt-runs.completed, alert.fired, recommendation.applied, recommendation.pushed.
- Personas regenerate now shows a diff of new vs old clusters with one-click apply.
- AI Agent Lens adds Amazonbot and Bytespider, bringing the tested bot count to twelve.
- Shopify integration: ecommerceinsights namespace metafields now respect Theme Editor visibility settings.
Direct API for Yotpo and Judge.me reviews. Brand voice URL crawl.
- Yotpo direct API integration. No more CSV exports for Yotpo customers; hourly refresh by default.
- Judge.me direct API integration. Same hourly refresh.
- Brand voice can now be derived from a public style-guide URL — paste, preview, edit, save.
- Image Enhancer: AI creative brief now writes against top three personas instead of brand average.
- Listing rewrite: capsule view shows the diff highlight against current value rather than a side-by-side replacement.
Workspace polish. ScrapingBee retry tuning. Internal bug fixes.
- Workspace-level theme polish: paper-grey palette refined for higher contrast on score badges.
- ScrapingBee retry budget tuned to halve fallback fetch time on Cloudflare-protected stores.
- Personas: minimum corpus warning ("under 50 reviews; clusters may be coarse") added when applicable.
- Schedule: per-task "Run now" buttons added next to each task entry.
AI Agent Lens, onboarding ring, email/Slack/webhook alerts, cron heartbeat.
- AI Agent Lens shipped. Paste any URL; get per-bot crawler verdict (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and friends), structured-data inventory, plain-text view an LLM extracts, AI Agent Readiness score, and auto-generated shopper test prompts with direct ChatGPT/Perplexity/Claude links.
- Onboarding checklist on Home: progressive 6-step ring (import → first audit → brand voice → reviews → first Monitor → first alert), auto-ticks from real state.
- Email alerts via Resend (RESEND_API_KEY) plus Slack plus webhook delivery wired into Alerts.deliver().
- Cron heartbeat at
/?api=cron_tickfor cron-as-a-service, plus "Run scheduler now" button on Schedule. - Hidden versions index page.
Sentiment Analysis shipped. Competitive Position promoted out of Beta.
- Sentiment Analysis: themes, voice-of-customer rollup, tone-gap detection between reviews and AI engine responses.
- Brand-wide Sentiment rollup with per-SKU table.
- Competitive Position moved out of Beta into the main Visibility section.
- SALES_READINESS.md audit doc.
Listing rewrite restored. ScrapingBee fallback. Beta hub. Action Inbox polish.
- Listing comprehensive rewrite restored. Amazon now gets capsule diffs plus bullets diff plus Enricher output stacked.
- PdpFetcher no longer caches blocked content. PdpParser handles array-wrapped JSON-LD plus unescaped control chars (the RobotShop fix).
- fetchText() falls back to ScrapingBee on 403 or Cloudflare challenge; finds llms.txt and robots.txt behind Cloudflare.
- Beta hub at
/?view=beta; single Beta link in nav replaces five separate entries. - Action Inbox polished: hero stats, filter pills, ranked cards, dismiss button.
- Alerts page redesigned with hero stats, fire cards, iOS-style toggles rules grid.
Analyze Page full audit detail. Crawler Access preflight. Force JS rendering.
- Analyze Page report now renders full audit detail via shared audit_report partial.
- Crawler Access preflight card on every audit (Score, Rufus, Analyze).
- "Force JavaScript rendering" toggle on Analyze (ScrapingBee with render_js).
- FUNCTION_AUDIT.md document.
Brand Voice. Monitor modes. Scheduler. Action Inbox home.
- Analyze Page report fix: full audit detail, not just buckets table.
- Brand Voice in Settings prefixes system prompt for all generation.
- Monitor: Neutral vs Persona-segmented run modes with cost preview.
- Scheduler plus frequency fields plus cron docs plus Run-history page.
- Action Inbox becomes the default home page.
Nav restructure. Channel-aware Listing. Analyze a Page. Image Enhancer pillars.
- Restructured nav: Visibility and Optimize headers.
- Unified Listing view: channel-aware rewrites (Rufus for Amazon, Enricher for D2C and retailer).
- Built Analyze a Page: paste-any-URL audit, channel detection by host.
- Image Enhancer 4 named pillars: Feature Sentiment, Mobile Glanceability, Competitive Benchmarking, AI Creative Brief.
- Wired Image Enhancer to Persona reviews.
Channel-resolved score. URL auto-redirect. Retailer banner.
- Single channel-resolved score nav entry: Amazon to Rufus, D2C/retailer to PDP.
- Auto-redirect score and rufus URLs based on SKU channel.
- Retailer banner plus recommendation filter (no robots.txt advice for Macy's SKUs).
- Products column score is channel-aware.
Settings page consolidation. Personas v2. Rufus 15-relation breakdown.
- Settings page with Catalog Import, Prompt Library, Personas, and Alert Rules cards.
- Personas v2: CSV upload plus LLM clustering (4-8 buyer personas from real reviews).
- Rufus / COSMO report clarity update: explicit 15-relation breakdown panel.
- Labels renamed: PDP Score, Prompt Runs, PDP Enrich, Rufus Score.
Common questions
How often does eCommerce Insights ship?
How are breaking changes handled?
Where is the RSS feed?
/docs/changelog/feed.xml. Updated when each version ships.Can I subscribe to release emails?
Ask AI about this page
Open this topic directly in your favourite AI assistant — the prompt is prefilled.
Try this yourself
See eCommerce Insights on your own catalog. The free trial covers your full SKU set across five AI engines.
Start free trial