Launching today
AgentKey — Agent Access Management
Access governance for AI agents. Free.
1 follower
Access governance for AI agents. Free.
1 follower
Your AI agents need API keys for GitHub, Stripe, Linear — today they live scattered in env files with zero governance. AgentKey lets agents request access, humans approve once, credentials get vended encrypted and every call is audited. Built on Vercel. Free forever.
Hey Product Hunt 👋 — maker here.
I built AgentKey because I got tired of pasting API keys into .env files every time I spun up a new agent. No record of which agent had access to what, no approval workflow, no audit trail — just hope.
AgentKey flips the model: agents start with zero access and earn it through explicit human approval.
Here's what's actually different:
🤖 Agents drive procurement, humans approve. An agent hits a missing tool, submits a suggestion with a reason. Multiple agents can back the same suggestion. You see the demand, decide once, and the whole fleet gets it. You're the approver, not the curator.
🔐 Credentials are vended, not stored. AES-256-GCM encryption at rest. Agents fetch on demand and never hold raw secrets.
📜 Every action is logged. Who requested what, who approved it, when credentials were fetched. One-click revoke.
🧠 AI-powered setup guides. Drop a docs URL, get step-by-step instructions for creating the right credential.
🔌 Framework-agnostic. Any agent that speaks HTTP works — Claude Code, Cursor, LangChain, CrewAI, custom.
Stack: Next.js, Neon, Clerk, Upstash, deployed on @Vercel .
Free forever at agentkey.dev, source-available (BSL 1.1 → Apache 2.0 in 2030), self-hostable.
I'll be in the comments all day. Ask me anything — architecture, the "agent-driven catalog" bet, why I chose BSL over MIT, or what's broken.