DEV Community

Vivek V.
Vivek V. Subscriber

Posted on

AgentBoard: A Discovery Platform for the Agentic AI Community

DEV Weekend Challenge: Community

This is a submission for the DEV Weekend Challenge: Community

The Community

I write for The Agentic Engineer, a weekly newsletter tracking the agentic AI space. The readers are builders. They ship autonomous agents, multi-agent workflows, and AI tools for a living.

The number one question I get: "Which framework should I use?"

There are 15+ major agent frameworks right now. New ones launch weekly. You either scroll GitHub trending, ask Twitter, or hope someone wrote a comparison post this month. There's no single place to browse, compare, and discover what's out there.

That's the community AgentBoard serves: agentic AI builders who need a front door to the tools they're building with.

What I Built

AgentBoard is an open directory for AI agents, frameworks, and tools. Think Product Hunt, but only for agentic AI.

You can:

  • Browse 15 pre-seeded agents across 8 categories (dev tools, automation, research, infrastructure, and more)
  • Search by name or filter by category
  • Sort by GitHub stars, upvotes, or newest
  • Click into any agent for the full breakdown: tech stack, creator, description, links
  • Upvote and bookmark your favorites (persists locally)
  • Submit your own agent through a simple form

No login. No backend. Everything runs client-side with localStorage.

Demo

Live site: natearcher-ai.github.io/agentboard

The home page shows community stats and featured agents. The Discover page is the core: real-time search, category filter tabs, and three sort modes. Click any card for the full agent profile. Hit Submit to add your own.

Code

GitHub logo natearcher-ai / agentboard

AgentBoard — Community-driven AI agent discovery platform. Browse, share, and celebrate AI agents built by the community.

AgentBoard 🤖

Community-driven AI agent discovery platform.

Browse, share, and celebrate the AI agents, skills, and tools shaping the future of autonomous AI.

🔗 Live: natearcher-ai.github.io/agentboard


What is AgentBoard?

AgentBoard is an open directory for the agentic AI community. Think Product Hunt meets Awesome Lists — specifically for AI agents and tools.

Features:

  • 🔍 Browse & search a curated directory of AI agents
  • 🏷️ Filter by category (dev tools, automation, research, etc.)
  • 📋 Detailed agent profiles with tech stack, links, and descriptions
  • ➕ Submit your own agents via a simple form
  • ⬆️ Upvote and bookmark your favorites
  • 📊 Community stats dashboard
  • 🌙 Beautiful dark theme with smooth animations
  • 📱 Fully responsive (mobile + desktop)

Screenshots

Home

The landing page features a hero section, community stats, and featured agents.

Discover

Filterable, searchable directory with category tabs and sort options.

Submit

Clean form to contribute your own AI agent to the directory.

How I Built It

Stack: React 19, Vite, Tailwind CSS v4, React Router, Lucide icons.

I went with a static SPA on purpose. No database means no hosting costs, no auth complexity, and anyone can fork it and run it locally in 30 seconds. The seed data lives in a single JS file. Adding a new agent is one object in an array.

The UI is dark-themed with a purple/blue gradient palette. Cards have hover animations. The discover page supports real-time search, category tabs, and three sort modes. The submit form validates inputs and drops you straight into the directory after submission.

Deployed via GitHub Actions to GitHub Pages. Push to main, site updates in under a minute.

MIT licensed. PRs welcome.

Top comments (2)

Collapse
 
meimakes profile image
Mei Park

The "which framework should I use" question is genuinely impossible to answer right now without spending a weekend on GitHub. Having a single directory with category filters and stars makes this immediately useful. Curious — are you planning community-contributed reviews or comparisons, or keeping it as a discovery layer?

Collapse
 
vivek-aws profile image
Vivek V.

Both, eventually. Discovery first. Reviews second.

Right now the directory solves the "what exists" problem. 15 agents across 8 categories, filterable, sortable by stars. That's the baseline.

Community reviews are next. I want builders rating frameworks they've actually shipped with, not drive-by opinions from people who read the README. One honest "I used CrewAI for 6 months and here's what broke" is worth more than 50 upvotes.

Side-by-side comparisons are on the roadmap too. Pick two frameworks, see tech stack, stars, use cases, and community sentiment in one view. No more opening 12 GitHub tabs to figure out if you want LangGraph or AutoGen.