Feedship AI Catalog Feeds for WooCommerce

Feedship AI Catalog Feeds is a free WooCommerce plugin that publishes your product catalog in the machine-readable formats used by AI shopping channels. It generates an OpenAI ACP product feed built to OpenAI’s published Commerce Feed Reference (JSONL & CSV), writes /llms.txt, /llms-full.txt and /llms-commerce.txt catalog files on your own server, serves ACP and UCP discovery manifests, and audits every product with a deterministic 0-100 AI Readiness Score. No coding required.

Everything runs on your own WordPress server. Feeds are generated on the schedule you choose – WordPress cron, your own cron URL, or a manual button – and written as static files. The plugin does not require external accounts, API keys, or a SaaS service.

The published files are plain text at public URLs on your site, so any reader that follows the relevant specification – an AI assistant, a crawler, or a feed validator – can fetch them like any other page.

Documentation | Installation | FAQ | Support

Feedship AI Catalog Feeds is an independent plugin: not affiliated with, endorsed by, or sponsored by OpenAI, Google, Anthropic, Perplexity, Shopify, or WooCommerce. All product names and trademarks belong to their respective owners and are used only to describe compatible protocols and destinations.

What the plugin publishes

• OpenAI ACP product feed (JSONL & CSV) – one row per purchasable variation with identifiers, images and policy fields, built to OpenAI’s published Commerce Feed Reference
• UCP discovery – a /.well-known/ucp manifest plus a UCP catalog search REST endpoint, per the Universal Commerce Protocol co-developed by Google and Shopify
• ACP discovery – a /.well-known/acp.json manifest
• The /llms.txt family – /llms.txt (store index), /llms-full.txt (full catalog with prices, stock and deep links) and /llms-commerce.txt (shipping, returns, payments), following the llmstxt.org format
• Product JSON-LD – hasMerchantReturnPolicy, shippingDetails and FAQPage markup added through the schema filters of Yoast SEO or Rank Math when they are active
• AI crawler monitor – reports whether GPTBot, OAI-SearchBot, ChatGPT-User, PerplexityBot, ClaudeBot, Google-Extended, CCBot and Amazonbot can reach your store

Setup in three steps

1. Run the setup wizard – Map your Brand and GTIN sources and set your seller name and return policy URL. The 1-click field mapper reads identifiers from WooCommerce core, Yoast SEO, Rank Math, Germanized, or any custom meta key.
2. Publish the files – The JSONL/CSV feeds, the /llms.txt family and the ACP/UCP manifests are generated automatically on your schedule.
3. Submit your feed URL – Copy the hosted JSONL/CSV URL into OpenAI merchant onboarding for ChatGPT Shopping if you want to list your products there.

That is the whole setup. No XML editing, no spreadsheets, no external dashboard.

Features

OpenAI feed generation. Every purchasable variation becomes its own feed row, linked to the parent group_id with a consistent variant_dict (color, size, material). The WebP/AVIF-to-JPEG/PNG auto-resolver converts images so image_url meets the feed spec’s JPEG/PNG requirement. Titles and descriptions are HTML-entity decoded with a short-description fallback, and output is plain UTF-8 without a BOM so strict CSV and JSONL parsers read every header and row.

AI Readiness Score (0-100). A deterministic 7-point audit – required fields, brand, image format, description depth, GTIN/MPN identifiers, attributes, FAQ coverage – rendered as colored badges in Products > All Products, grouped into AI Ready / Needs Polish / AI Blocked, with one-click bulk recalculation. The same product data always produces the same score.

1-click product identity mapping. Pull GTIN, UPC, EAN, ISBN, MPN, and Brand from WooCommerce core, Yoast, Rank Math, Germanized, or any custom meta key, with Modulo-10 GTIN check-digit validation and a store-wide fallback brand.

The /llms.txt family, e-commerce flavored. /llms.txt (short store index), /llms-full.txt (every product with SKU, price, stock, and deep links), /llms-commerce.txt (shipping, returns, payments). Detects Yoast SEO or Rank Math and extends their output via native filters instead of duplicating files.

Dual-protocol discovery. /.well-known/acp.json (OpenAI ACP) and /.well-known/ucp (Google & Shopify’s UCP) manifests with automatic merchant logo resolution – Customizer logo, Site Icon fallback, or manual override.

Per-product control. A dedicated metabox on every product: readiness checks, brand and availability overrides, condition/gender/age-group fields, an exclude-from-all-feeds toggle, and conversational Q&A published in /llms-full.txt and as FAQPage JSON-LD.

AI crawler guard. Scans your effective robots.txt output, live-probes /llms.txt and your feed URL, and shows per-bot status for eight AI crawlers, so you can see which requests would be blocked before you submit your feed.

Generated on a schedule. Feeds are generated on the schedule you choose – WordPress cron, your own cron URL, or a manual button – not during shopper page loads. They are written atomically as static files and served straight from disk; the plugin adds no front-end scripts or styles, and generation never runs inside a shopper’s request.

Self-hosted, end to end. Everything is generated on your own server under wp-content/uploads/feedship-feeds/ with directory listing disabled. The plugin never sends your catalog to a third-party service – its only HTTP calls are loopback checks to your own site – and an optional clean uninstall removes every stored trace. The published feed and discovery URLs are public by design: that is how feed readers fetch them.

Developer friendly. WordPress hooks and filters throughout (like feedship_merchant_logo), catalog reads through WooCommerce CRUD APIs, and declared HPOS + Cart/Checkout Blocks compatibility.