Hi Product Hunt! 👋 Three weeks ago I lost my job - the client's budget got cut three days before my probation ended. I started job hunting the modern way: with AI agents doing the heavy lifting. But every tracker I tried was built for humans clicking forms, while my actual workflow was "Claude, I applied to Bosch as Tech Lead, prio A, follow up Friday." So I built jobctl - a tracker where the agent is the first-class user. It speaks MCP (remote server, OAuth 2.1, official MCP Registry), so any agent can add applications from a single sentence, log interviews with attendees and prep links, attach tailored CVs it just generated, and answer "what's burning today?" I read the dashboard; the agents do the bookkeeping. The meta part: jobctl was also built by the same agentic workflow it serves - spec-first development, deterministic quality gates, AI agents writing the code. And it's not a demo: I'm running my own active job search through it right now - every application, interview and follow-up in my tracker was entered by an agent while I focused on the actual interviews. It's free while I figure out where this goes. I'd love your feedback - especially from fellow MCP builders: tool design, OAuth flow, what you'd want your agent to track next.

jobctl

The job application tracker your AI agent keeps updated

The job application tracker your AI agent keeps updated