Optionality over lock-in
We optimize for reversibility, not for predicting one vendor's roadmap. The runtime speaks Claude, OpenAI, and OpenCode through one tool layer, so you bring the models you trust and swap them without a rewrite.
About
HQ brings decision intelligence to your Slack and Microsoft Teams. @mention a specialist and it does not just answer questions, it researches, decides, writes, files documents, sends email, and builds, each inside its own secure machine, with memory of your work and a record of everything it does. We built the whole platform, top to bottom, to make that dependable enough for real work.
The company
HQ is built by The Intelligence Company AB (publ), or TIC, based in Stockholm, Sweden. We have spent years building platforms for bringing in, and acting on, very large amounts of data, with agents at the center of how that data gets used.
HQ began as infrastructure we built for our own demanding work. When it became clear how much of it, the data ingestion, the agent-first design, the safety and accountability, other teams needed too, we turned it into a product anyone can pick up.
Our research arm, Truespar, releases several of the systems beneath HQ and TIC, like Traverse and Torque, as standalone products you can download and run yourself.
Trust
Data protection is in the foundations, not the fine print. We come from years of handling sensitive data under real scrutiny, and it shows in the architecture: choose EU or US data residency across two fully separated and isolated datacenters, governance designed for the EU AI Act, and an audit-first identity model where every agent action carries an identity chain and lands in a record that cannot be edited after the fact.
Under the hood
HQ runs on qOS, the platform we built for agents to live on. Every conversation boots its own isolated machine in a fraction of a second, over a data layer we wrote ourselves: search, cache, storage, graph, browser, email, documents, extraction.
Several of these are the same systems Truespar ships on their own, so the foundation under HQ is more proven than a new product usually gets to stand on. It is the part you never see, and the reason the part you do see can be fast, private, and dependable.
What we believe
We optimize for reversibility, not for predicting one vendor's roadmap. The runtime speaks Claude, OpenAI, and OpenCode through one tool layer, so you bring the models you trust and swap them without a rewrite.
Every action an agent takes is tied to the person who asked, and written to a record that cannot be edited after the fact. Built in from the start, not bolted on later.
Data residency you choose, governance designed for the EU AI Act, and privacy in the architecture rather than the disclaimer.
We would rather show a real capability than a slogan. When something is a goal rather than a fact, we say so.
HQ is currently invite-only while we onboard teams. Request access and start in your own Slack or Teams.
Request access