ReplayAutonomous app testing for teams who ship at AI-speed
Promoted
Prompteneering
The largest free library of expert AI agents
8 followers
The largest free library of expert AI agents
8 followers
The largest free AI agent library — 280,117 expert AI agents across 65 fields and 20 languages. Each is a structured system prompt tuned to the right tool: Claude, ChatGPT, Cursor, Midjourney, Veo, ElevenLabs, Zapier, and more. Discover specialists, assemble a team, and deploy the whole crew as a bundle, files, or share link. Free public API and MCP server. No login.
Hey Product Hunt 👋
I kept hitting the same wall with AI: every serious task meant re-writing the same kind of prompt from scratch (role, context, constraints, output format), and the good ones I wrote got lost in a chat history I'd never find again. There was no library, no standard, no quality bar.
So I built Prompteneering.
It's a network of 23,468 expert AI agents across 63 professional fields, and they're not throwaway prompts. Every agent follows a strict framework (Role, Context, Task, Constraints, Output format), is tuned to the right tool (Claude, ChatGPT, Cursor, Midjourney, Veo, ElevenLabs, Zapier, not just chatbots), and is built for real, paid work.
The part I'm most proud of: you don't just grab one agent, you build a team. Add the specialists you want, give them names and job titles, then deploy the whole crew as a combined system prompt, individual files (Claude Projects / Custom GPTs), or a share link. No account needed.
It's also in 9 languages, tagged for UK / EU / US markets where the rules differ, and there's API + MCP access for teams who want it wired into their stack.
It's free to explore. Drag the brain, click a region to dive in. Would love your feedback, and the agents you'd want added next 🙏 or just add your own
Free at 👉 https://prompteneering.com
Geoff - Launchcore
A few people asked "okay, but how do I actually use 23,000 agents?" — fair question. Two things I didn't shout about enough:
1. You search in plain English. Don't think in keywords. Type "write a cold email that handles pricing objections" or "kubernetes blue-green deploy" and it ranks the right specialist first. No hunting through a directory.
2. It plugs into your stack. There's a free MCP server — connect it to Claude, Cursor or Claude Code and your AI can pull any agent (or assemble a whole team) without leaving your editor. It'll even orchestrate that team as a loop — research → build → review, running until the job's actually done.
So it's less a prompt library, more an agent layer you wire into the way you already work.
Drop your field in the comments 👇 and I'll tell you the 3 agents I'd start with — or add a new one if we're thin there 🙏
Subanana
This sounds really interesting @geoffbosley The agent team builder is a clever idea! most prompt tools stop at the individual level. I wish they had something liek this within the tool by default.
Congrats on the launch
@lakshya_singh Thank you I agree. We are moving into loops and teams of agents not just one. So we expect all our agents to work concurrently.