About CurateAndHarnessBot

Information for website operators · Updated September 20, 2026

CurateAndHarnessBot is operated by Curate & Harness to help subscribers discover clothing from selected brands. Recommendations link readers to the merchant’s own product pages to shop.

What the bot fetches

The bot reads public Shopify product-catalog JSON, including product titles, prices, categories, colors, sizes, availability, and image URLs. Requests are limited to brands configured for our clothing-search tool. It does not access customer accounts, order records, or merchant administration, and it does not add items to carts or complete purchases.

How the content is used

We filter available products and match them to a subscriber’s stated clothing preferences. Selected product attributes are sent to the OpenAI API for ranking and short explanations; product images are not sent to the AI model. Results are displayed to subscribers and may be emailed through Resend.

Catalog requests are triggered by searches, and successful catalog results are cached for 30 minutes. A refresh can request up to 20 pages of 250 products each, excluding retries. Request volume varies with searches and catalog size.

Bot identity

User-Agent
CurateAndHarnessBot/1.0 (+https://curateandharness.com)

Shopify requests use Ed25519 Web Bot Auth signatures with Signature-Agent: "https://curateandharness.com". Our public signing-key directory is available without authentication.

Robots.txt and crawl controls

Robots.txt and site terms are reviewed when brands are configured. The current implementation does not automatically recheck robots.txt or enforce crawl-delay directives. Website operators should contact us directly about exclusions, changed policies, or request-volume concerns.

Opt out or contact us

To request that we stop fetching or displaying content from your website, email [email protected] with the subject “CurateAndHarnessBot opt-out.” Include your domain, any specific URLs or sections concerned, and your relationship to the website.

Opt-out requests are handled manually; sending an email does not automatically change the crawler’s configuration. Use the same address for questions about the bot, corrections, or access concerns.