CatalogPulse
An open-source AI e-commerce linter that audits the AI discoverability of product pages and provides structured-data remediation code through deterministic DOM parsing, real-time market signals and open-weight model auditing.
Timeline (last 90 days)
CatalogPulse is a zero-install, open-source AI e-commerce linter where merchants enter a product URL and within 10 seconds get a discoverability score, a defect list, and ready-to-paste Schema.org Product JSON-LD fix code.
CatalogPulse uses a three-layer processing pipeline: the first layer uses Cheerio for deterministic DOM parsing, the second layer calls SerpApi to obtain Google Shopping market signals, and the third layer outputs structured JSON based on the open-weight Google Gemma 2 (gemma-2-9b-it) model.
There are three reasons CatalogPulse's technical choice uses an open-weight model: on cost, marginal cost is close to zero; on data sovereignty, customer catalog data does not have to pass through a third-party cloud service; on controllability, it avoids closed-source models silently changing behavior or endpoints being deprecated.
The author tested CatalogPulse with an independent D2C shoe and apparel store, and the flagship product page scored 65 (grade C); the store actually sells 6 sizes, but the Shopify theme only output the default variant in the root Offer node and did not declare the gtin13 attribute; the SerpApi panel showed a secondhand reseller on Poshmark listing the same item at half price.
Pasting the JSON-LD code block generated by CatalogPulse into the theme header made the page pass validation within two minutes, with complete Brand, SKU, and Offer attributes, without hiring a $500-per-month SEO consultant.
The CatalogPulse project roadmap includes crawling the entire sitemap.xml every Sunday night and sending email notifications when theme updates break the structured data.
All Facts (6)
CatalogPulse uses a three-layer processing pipeline: the first layer uses Cheerio for deterministic DOM parsing, the second layer calls SerpApi to obtain Google Shopping market signals, and the third layer outputs structured JSON based on the open-weight Google Gemma 2 (gemma-2-9b-it) model.
50%UnverifiedThere are three reasons CatalogPulse's technical choice uses an open-weight model: on cost, marginal cost is close to zero; on data sovereignty, customer catalog data does not have to pass through a third-party cloud service; on controllability, it avoids closed-source models silently changing behavior or endpoints being deprecated.
50%UnverifiedCatalogPulse is a zero-install, open-source AI e-commerce linter where merchants enter a product URL and within 10 seconds get a discoverability score, a defect list, and ready-to-paste Schema.org Product JSON-LD fix code.
50%UnverifiedThe author tested CatalogPulse with an independent D2C shoe and apparel store, and the flagship product page scored 65 (grade C); the store actually sells 6 sizes, but the Shopify theme only output the default variant in the root Offer node and did not declare the gtin13 attribute; the SerpApi panel showed a secondhand reseller on Poshmark listing the same item at half price.
50%UnverifiedPasting the JSON-LD code block generated by CatalogPulse into the theme header made the page pass validation within two minutes, with complete Brand, SKU, and Offer attributes, without hiring a $500-per-month SEO consultant.
50%UnverifiedThe CatalogPulse project roadmap includes crawling the entire sitemap.xml every Sunday night and sending email notifications when theme updates break the structured data.
50%