A reference library of real product interfaces, cut into labelled sections and wired into an MCP server, so Cursor or Claude can pull a working build prompt instead of you eyeballing a screenshot and guessing at the CSS.
Table of Contents
Every designer has done this: find a website with exactly the hero section you want, zoom in, squint, and try to recreate it from memory. UI Discovery skips the squinting. It has already captured 2,647 websites, 62 web apps and 3,836 iOS apps, cut them into 111,083 individually labelled sections, heroes, pricing tables, footers, onboarding screens, and attached each one a color palette, a font stack and an AI-ready build prompt someone can hand directly to a coding agent.
A library built for agents, not just browsing
Most design-inspiration sites assume a human is scrolling. UI Discovery assumes an AI tool might be the one asking. It runs a hosted MCP server with a documented set of tools, `search_sections`, `search_screens`, `search_flows`, `search_sites`, `compare_references`, `find_similar`, `get_design_tokens`, plus Pro-gated tools for a full design brief and an image-based build plan, installable into Cursor, Claude Code or Codex with a single command. A parallel REST API backs the same data for anyone who’d rather call it directly, with documented pagination and rate limits rather than an undocumented scraper’s nightmare.
That’s a meaningfully different product shape from a Pinterest-style mood board. A mood board gives you a saved image to look at. UI Discovery gives you a structured prompt, already carrying the layout, colors and type an agent needs to rebuild the pattern rather than merely admire it.
Pricing, and what a capture comes with
Browsing is free; the deeper tools sit behind Pro at $15 a month and Team at $39 a month. Each captured section carries an extracted palette and font information alongside the build prompt, and the library distinguishes its own captures from imported third-party reference archives, labelling which is which on the relevant pages rather than blurring the line. For a product whose entire pitch rests on being an accurate, labelled reference, that kind of self-disclosed provenance matters more than it would for a typical inspiration gallery.
A small operation behind a large library
One detail worth flagging for anyone building a workflow around this: the service’s own terms and privacy pages list a Gmail address, not a company domain, as the point of contact, and there’s a stated content-removal process for anyone who wants their captured site taken down. That’s a reasonable, informal process for a reference library, but it’s also a signal of scale worth weighing against the size of the dataset itself. On the privacy side, the stated policy is that daily visitor counts come from a same-day salted hash rather than stored IP addresses, and the operator says it doesn’t sell personal data or use it for advertising.
None of that changes what the tool does well. It just means the honest way to use 111,083 screenshots of other people’s finished work is to treat it as a reference for layout and pattern, the same way designers have always borrowed structure from sites they admire. It isn’t a shortcut past the part where someone still has to make deliberate choices about the product being built.
That distinction matters more now than it did a few years ago, because the output of a tool like this no longer stops at a mood board a human interprets loosely. It feeds directly into an agent that will happily execute the structure literally, gaps in taste and all, unless a person is still making the calls about what actually belongs in the finished product.
Author:
Jan Bielik
CEO & Founder of Webiano Digital & Marketing Agency
Sources
- UI Discovery — agentsindex.ai – capture counts, MCP tool set, API details, pricing, provenance labelling
- UI Discovery — agentsindex.ai (data sheet) – contact details, privacy policy on visitor counts, content-removal process
- UI Discovery — mcpservers.org – MCP server installation, what each reference includes
| Citing this article? Brief excerpts are welcome. Please credit Webiano.digital, name the author where stated, and include a link to https://webiano.digital and to this original article. Full or substantial republication requires prior written permission. Read our Copyright and Content Use Policy. |
This article was prepared with the assistance of artificial intelligence tools. The content underwent expert human review, and Webiano Digital & Marketing Agency assumes editorial responsibility for its final version and publication.
