HQ is built on a simple observation: when everyone uses AI individually, the gains stay individual. One person finds a prompt that works, it lives in their chat history, and the rest of the team starts from zero the next day. HQ is the shared layer underneath — a company brain that holds knowledge, skills, and workflows so that AI gets sharper the more the team uses it, and each person's breakthrough becomes everyone's baseline.
The mechanics are file-based and command-driven. Your active working directory in Claude Code, Codex, or Cursor points at the shared HQ context, and ChatGPT or Claude chat connects through MCP. Any prompt or workflow can be saved as a reusable slash command, and /hq-sync pushes it to the team. Knowledge — playbooks, brand guidelines, onboarding docs, competitive research — accumulates alongside them, so new projects inherit what the company already knows instead of onboarding from scratch.
Around that core, HQ adds the pieces teams usually stitch together themselves. HQ Bots bring persistent AI teammates into Slack and email that work from your skills and knowledge. HQ Deploy turns internal dashboards and tools into live links in a single command, with public, password-protected, or company-only access. A secrets manager stores API keys and credentials and injects them at runtime so they are never pasted around. HQ Cloud handles sync and team management, and the core is open source. Teams can start free with $10 of usage included monthly, and because the context layer is separate from the models, swapping AI vendors does not reset your setup.
