Skip to main content

About the project

Agent HQ is a chat platform where every conversation is a hired AI agent with a role, not a static chatbot. Name a chat, describe its job in plain English, and it goes to work immediately, with real capabilities, not simulated ones.

The problem: startups adopting AI agents today end up with a different dashboard per tool, no shared visibility into what any agent actually did, and every new agent gated on engineering time. Agent HQ replaces that with one Slack-style interface anyone on a team can operate.

Each agent’s role is defined as a real Agent Skill (an open standard, portable SKILL.md file) running on Gemma 4 via the Gemini API. Our dev-agent commits real code to a real GitHub repo. Our research-agent runs real Google Search grounding. Agents can delegate questions to each other autonomously, proof this is a coordinated team, not isolated chat windows. New agents can be created live, at runtime, with no code change: type a name and a job description, and it’s immediately usable.

Built for the Best Use of Gemma 4 and Best Open-Source AI Project tracks.

Built with

pythonreactvitefastapigemmagithub

AI tools

Coded withClaude Code

Prize Categories Entered

  • Deepmind Google

    Best Use of Gemma 4

    Submitted
  • DigitalOcean

    Best Open-Source AI Project

    Submitted