I gave my AI access to 600,000 real app & web screens. Here's the workflow: 1. Connect @mobbin MCP to your AI (I used Claude, works w/ Codex, Cursor, Antigravity too) 2. Ask it to research how top apps in your industry design what you're building (eg. fintech onboarding) 3. Ask it to compile the findings into a visual report 4. Build, but give it context first: Attach the report, a spec or PRD, your design system the result - without the MCP: generic output - with the MCP: Claude picks up the industry context, sees the end-to-end flow, considers the edge cases. Think of Mobbin MCP as a librarian, not a designer. It hands your agent a library card to 600K+ real screens, full flows and animations from products already shipped. Your agent sees what good looks like, then gives you a starting point to build on. here's the full workflow + detailed prompt 🔍 'mobbin' on youtube

I studied 𝟮,𝟏𝟎𝟖 𝗱𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱𝘀 across 𝟱𝟯𝟲 apps & websites. most dashboards work, but they're forgettable. I didn't want to make another "how to design a dashboard" video, so i went down a rabbit hole to find out what makes them stick. @mobbin's data: → 19% apps/websites use the dashboard as their home screen → AI dashboard screens get about 2× more saves & exports this 5-minute video breaks down the patterns behind dashboards that people actually remember, from @Shopify, @Apple, @runna etc 🔎 'mobbin' on youtube

I studied 𝟭,𝟰𝟲𝟬 𝗼𝗻𝗯𝗼𝗮𝗿𝗱𝗶𝗻𝗴 𝗳𝗹𝗼𝘄𝘀 across 𝟵𝟴𝟲 apps & websites. the average app has 25 onboarding screens. and some of the longest ones are the most successful. so it's probably not about making onboarding shorter. then what actually makes a good onboarding flow? @mobbin's data: → 22% of apps personalize during onboarding (AI apps: only 7%) → 22% show a paywall during onboarding → 6.3% do both This 10-minute video breaks down how top apps design onboarding, from @duolingo, @Headspace , Bump by @amoamoamo, @Airbnb, and more 🔎 'mobbin onboarding' on youtube