Independent research · StatusBird LLC

The AI-Built Store Index

How many web stores built with Lovable, Bolt, v0, Replit and similar tools exist in public sources, and what search engines, AI assistants and shoppers actually receive from them. Aggregates only. No store is named.

First edition in preparation. Candidate sites are being gathered from certificate transparency logs and public code repositories, fetched once the way a crawler does, and rendered once the way a shopper does. Counts are published here when the first pass completes. Meanwhile the free store check runs the same measurements on any single store.

Method

Where the sites come from: Certificate transparency logs for builder-hosted domains, and public code repositories that a builder synced and that contain commerce code (cart, checkout, payment provider, product). A repository's published site is taken from its homepage field or README, or inferred from its name and verified by fetching. Common Crawl's URL index is added when it answers. This finds a sample, not the population; it is biased toward stores whose owners publish code or use a builder-hosted address.
What counts as a store: The rendered homepage shows a cart, checkout or buy control, or carries Product structured data, or shows three or more prices on a site whose code contains commerce logic. Placeholder projects (builder welcome pages) are excluded.
What a crawler receives: One plain HTTP fetch of the homepage with an identified user agent, no JavaScript. Words are counted after removing scripts, styles and markup. Fewer than 60 words with an application root element counts as an empty shell. Titles matching a builder's default (for example "Vite App") count as default.
What a shopper sees: One load in headless Chromium with a settle period: rendered text, Product structured data, images without alt text, axe-core WCAG 2.1 A and AA failures (critical and serious), largest contentful paint, JavaScript errors, trackers requested before any interaction.
Credential patterns: Up to three of a site's own JavaScript files are matched against documented key prefixes (Stripe secret, AWS, SendGrid, OpenAI, Anthropic, private key blocks) and JSON web tokens are decoded only to read their role claim. Values are never stored. No database, API or data endpoint belonging to any site is ever queried.
What is published: Aggregates only. No store is named on this page or in the data file. A store owner who wants their own measurements can run the free check. Builders named here are identified from public page markers; StatusBird LLC is not affiliated with any of them.
Refresh: Monthly. Each edition re-fetches every site so the trend is like for like.

Questions, corrections, or a builder's response: studio@statusbird.io. Responses from builders are published verbatim alongside the next edition.