Privacy
Two different things: what we collect from visitors to this site, and what we hold on behalf of customers. If anything here is unclear, write to toivo@stuhi.org and we will answer plainly.
If you join the waitlist: your email address, and your company name if you choose to give it. Nothing else, and nothing is required beyond the address.
If you sign a letter of intent: the name, role, company and email you type into that form, together with what you wrote and the time you submitted it.
We hold the context you give us: the Slack export, documents or text you upload, and whatever is in it — including the names of your people and what they wrote. Nothing is collected automatically. You choose what to hand over, and you can hand over less.
The Slack app reads only the channels you invite it into. A bot sees history for channels it is a member of, not your whole workspace. It is never in a private channel unless someone adds it, it never sees direct messages between employees, and removing it from a channel ends its access immediately.
We do not use your content to train models, sell it, or use it for anything other than running VANAV for you.
We count page views using Vercel Analytics, which is cookieless. It sets no tracking identifier, does not follow you between sites, and does not build a profile of you.
We deliberately do not use Google Analytics. Several European data protection authorities have found its transfers unlawful, and using it would contradict everything else on this page.
No advertising scripts, no third-party trackers, and nothing that writes a cookie.
Only to contact you about VANAV, to tell you when a slot opens, or to reply to something you sent us. We do not sell it, share it with advertisers, or use it to train anything.
We keep it until you ask us not to, or until it is obviously stale. A waitlist address from a company that never replied is not worth keeping and we will not.
Content and account records are stored in a Postgres database in Stockholm, along with any files you upload. The site itself runs on Vercel.
To answer a question, the relevant context is sent to a large language model. Today that is the Anthropic API, which processes it in the United States. We are moving this to Amazon Bedrock in an EU region so that it stays inside the EEA, and this page will say so once it has. Neither provider trains on content sent through their API.
The full list of processors we rely on — Vercel, Supabase, Anthropic, AWS, Slack, Stripe and Resend — and what each one does, is available on request.
Email toivo@stuhi.org and ask. You do not need to explain why, and we will confirm when it is done. You can also ask for a copy of what we hold.
For a company, deletion removes the context you gave us, any files you uploaded, and your account record, in one operation. We aim to complete a request within 30 days and in practice within a few working days. One honest limit: a running server may hold content in memory until it restarts, so we redeploy when immediacy matters.
Everything in the product demo, the company, the fourteen people and every message, is invented. It is not any real company’s data.
If you paste your own messages into the ingest page, they are processed to build a role and are not shared with anyone. Do not paste anything you would not be comfortable sending us in an email.