ForU Grid
Time Frame
2025
Role
UI/UX Designer
Platform
Web
Introduction
ForU Grid is an AI agent marketplace built within the ForU ecosystem — a place where users can discover, browse, and deploy AI agents for specific tasks. As the in-house Product Designer at ForU, I designed Grid from scratch: the marketplace interface, agent detail pages, category browsing, and the overall visual system that made Grid feel like a natural extension of the ForU platform.
Background
ForU was expanding beyond its reputation layer. The team wanted to give users a way to access and deploy AI agents directly within the platform. Grid was conceived as the marketplace layer: a dedicated space where agents could be listed, browsed by category, and launched with one click. The design challenge was making an AI marketplace feel approachable and trustworthy to a Web3-native audience without alienating users who weren't deeply technical.
Team
Collaborated with one other designer on this project, working alongside the ForU product and engineering team. Grid was a self-contained design initiative. We shared ownership across information architecture and final UI, with direction from the CEO and product lead.
Defining the Problem
How might we design an AI agent marketplace that feels intuitive and trustworthy to Web3 users — making it easy to discover, evaluate, and deploy agents without technical friction?
Process
Audit existing patterns → Define IA → Design system → Component build → Iteration.
I started by auditing existing AI and Web3 marketplaces to understand how browsability, trust signals, and agent categorisation typically work. I then mapped out the IA for Grid: homepage with featured agents, category filters, agent cards, and a detail page with usage stats, reviews, and a deploy CTA. The visual system was kept consistent with ForU's dark glassmorphic aesthetic while adding marketplace-specific patterns.
Key Design Decisions
Agent cards were designed to surface the most decision-relevant info at a glance: agent name, category tag, creator, usage count, and a one-line capability summary. The detail page follows a product-style layout — hero description, key capabilities, compatibility info, and social proof (ratings, deploys). Category browsing was kept flat and scannable rather than nested, reducing the number of clicks to find a relevant agent. The dark glassmorphic visual language from ForU's main platform was carried through Grid to keep brand consistency, with accent colours used to differentiate agent categories.
Research and Ideation
Referenced established AI marketplaces — GPT Store, Hugging Face Spaces, Wordware — to understand how agent discoverability and trust signals are typically handled. Web3-specific references included Lens Protocol and Mirror for aesthetic and UX patterns. Key insight: users in AI marketplaces need to understand capability before they commit to deploying — so the agent detail page design prioritised clear capability descriptions and real usage data over visual flair.
Result
Browse Agents
Agent Marketplace
Full Screens Overview
Agent Detail Page
Challenge and Most Interesting
The hardest part was designing for a product category that didn't fully exist yet in the ForU ecosystem — agents weren't live, so there was no real data to reference. The visual had to make Grid feel real and populated without being misleading. Balancing the Web3 aesthetic (dark, glassmorphic, dense) with marketplace usability requirements (scannable, low friction, quick to evaluate) was the central tension throughout the project.
Key Takeaways
Designing a marketplace from scratch requires getting the IA right before touching visuals — the browsing logic determines everything downstream. Web3 products don't have to look intimidating; clarity and trust signals matter more than visual complexity. Designing with placeholder data still requires the visual to feel credible, fake numbers and names need to feel plausible, not random. Owning a product end-to-end forces better prioritisation decisions than working in a larger team.