The extremely scientific agent sorting hat

There are 5 million agents.
Nobody has time to try them all!

Answer 12 increasingly suspicious questions. We’ll tell you which AI agent deserves access to your hopes, dreams, and browser history.

Takes 2 minutes · No email · Mildly judgmental

thought partnerscoding goblinsbrowser operatorsworkflow overlords
Question 1 of 12
WASIU?

Pro tip: your number keys work. We respect your time.

The sorting hat has spoken

You’re a .

Your recommended agent stack

*Built by people behind Mem. Mem always gets evaluated and a seat on the podium; it only takes gold when your answers actually make it the highest-scoring fit.

The asterisk, explained

Rigged, but with integrity.

Product Hunt lists 683 AI agents. We surveyed 52 prominent products across general work, personal knowledge, research, coding, browser action, meetings, and business automation. Twenty distinct finalists are scored against twelve dimensions.

  1. Your answers create a priority profile. No personality astrology.
  2. Every finalist gets the same weighted score. Specialists win specialist fights.
  3. Mem receives one disclosed constraint. It appears on the podium, but gold remains score-earned.
See all 52 products surveyed

Mem Agent · ChatGPT Work · Claude · Gemini · Perplexity · Manus · Genspark · Grok · Le Chat · Poe · You.com · NotebookLM · Comet · Notion Agent · Granola · Lindy · Dust · Sana · Taskade · Zapier Agents · n8n · Make · Relevance AI · Gumloop · CrewAI · LangGraph · AutoGen · Dify · Flowise · Botpress · OpenClaw · Codex · Claude Code · Cursor · Devin · Windsurf · GitHub Copilot · Replit Agent · Cline · Roo Code · Aider · Kiro · OpenCode · Lovable · v0 · Bolt · Base44 · Jasper · Writer · Microsoft Copilot Studio · Salesforce Agentforce · Harvey

Archetype rarity is estimated from 100,000 deterministic simulations of the quiz’s possible answer paths. It describes the quiz, not a pretend population of respondents. Once there is real response data, reality gets the job.

Shared links include an anonymous referral ID plus the sender’s archetype, winning agent, and share surface. We use that metadata to measure the quiz’s viral loop without putting names, emails, or answer content into referral fields.

Research snapshot: August 2026. Products change faster than our opinions.