Mockly
Design inspiration platform for the AI age - blending automated UI/UX discovery pipelines with a modern, search-driven design experience.
CLIENT
Mockly
SERVICE PROVIDED
Design, Development, GTM
TIMELINE
6 Months
Mockly is a design inspiration platform built for product designers. It lets users explore real-world UI from apps and websites to gather inspiration for their next projects.
Our team partnered with Mockly to design and engineer not just a platform, but an autonomous AI system that continuously curates, indexes, and understands digital interfaces from across the web - making Mockly one of the fastest-growing UI/UX libraries in the world.
A living, self-evolving library that could automatically discover new interfaces, describe them intelligently, and make them searchable through natural, design-centric queries.
Product teams lacked a navigable, shared reference for modern interface inspiration: existent platforms are littered with unrealistic experiments, design libraries are years out of date and keyword search missed the nuance of layout and purpose.
The team at Mockly set out to build a search experience that understands how designers think - queries like “card-based checkout” should surface comparable, high-quality references fast. At the same time, the library needed an autonomous curator to keep itself current without manual effort.
Autonomous UI Discovery (Automator) - Agentic crawlers explore real products and websites, behave like power users (log in, progress flows, detect state changes), and capture screens at the right moments without getting stuck or looping. The crawlers' mandate is clean, reliable acquisition only; curation happens downstream.
AI-Generated Understanding (Platform) - An LLM pipeline analyzes each screenshot to produce structured descriptors: layout, hierarchy, components, flow purpose, and brand cues. We normalize terminology to keep semantics stable as new intake arrives.
Vector-Based Smart Search (Platform) - Content is embedded and indexed in Azure AI Search with hybrid retrieval (vector + keyword) and semantic ranking. The system is tuned for intent-led queries (e.g., “clean fintech dashboards,” “social onboarding flows”) under tight latency and at growing scale.
Quality, De-duplication, and Ops (Platform) - Ingestion applies strict de-duplication, resolution checks, and safety screening so results remain trustworthy as volume increases. Relevance tuning and analytics protect precision as the catalogue expands.
Seamless UX - The front end prioritizes speed, clarity, and creative flow, making it easy to browse, compare, and save references while the automator keeps the library fresh in the background.
Mockly today serves hundreds of passionate users around the world and we're proud of what we managed to acheive together with the wonderful team at Mockly.
