Models, agents and rankings in one place — a fusion of what people go to Hugging Face, GitHub and Artificial Analysis for, built so you can run the work yourself.
Honest note: HQ is a free beta today. $19 a month is the planned price when paid plans switch on. You can sign in now with Google or any email, and bring your own keys or local models. See what's included
Each door opens a page that already exists. No made-up scores.
These are screenshots of the public HQ site, taken on Sat Nov 21, 2026. The product page uses a sample home screen with made-up names; your own Home shows only your numbers.
Checked against the site itself on Sat Nov 21, 2026. If a number is not here, we have not verified it.
Bring your own API keys or run models on your own computer. HQ does not require a new AI subscription.
The planned price for one HQ workspace. Use the models and apps you already pay for, or ones that run locally.
Paid plans are not switched on yet. HQ’s own site still says the beta is free. You can start today at no charge. When billing goes live, the price we intend to charge is $19/month. We will not take a card on this marketing site.
No. Paid plans are not switched on yet. HQ is a free beta: sign in at hq.abuz8ai.com/start, and you get a private workspace at no charge. $19/month is the price we plan to charge when billing is live. This page will be updated when that changes.
A private workspace where the AI tools you already use can work as a team: chat, approvals, a run log, and a spending meter. You bring your own API keys or local models. The exact setup steps live on the Connect page.
No. HQ is built so you can use the providers you already pay for, or models that run on your own computer. You may still pay those providers separately. HQ itself is free during the beta.
No. The 159 tools on /tools stay up, and their old URLs keep working. They are just no longer the thing this homepage sells. The same is true of guides, the store, downloads and other pages: they are hidden from this homepage and this nav, not deleted.
The /models and /leaderboards pages cite a source and a fetch date for every figure. Quality comes from the official SWE-bench leaderboard JSON; price and context come from the OpenRouter models API. Speed and Artificial Analysis scores are labeled unknown in the current snapshot (we link to Artificial Analysis rather than copying scores we are not licensed to republish). We do not invent benchmark numbers.
ABUZ8 LLC, an Ohio company. Email support@abuz8ai.com only. We do not list a phone number. Privacy and Terms apply. Refund rules for anything you later pay for are on the refund policy.