Privacy policy
Last updated 26 October 2026
Quillm is run by Streamfog Inc. (“we”). This page says what we keep, where, and who else sees it.
What we store
Your name and email address from Google sign-in; your plan, and if you pay, Stripe's IDs for you and your subscription; the workspaces you belong to and their members; the pages and datasets your agents create, every revision of them, and a log of what each agent and person changed; values people save inside a page; a record of every tool call an agent makes (which tool, when, whether it worked, and for failed calls the error and the first part of the input), used to fix problems; how many times a day each page is opened and by whom (a member's email, or for share links and the example workspace a code made from the visitor's address and browser that changes every day, so nobody can be recognised from one day to the next); and, from your first sign-in, how you found Quillm (the website that linked you here, a tag in that link, and the first page you saw); which emails we sent you, and whether you turned them off.
All of it lives in one database on a server run by Fly.io in the United States, copied continuously to Cloudflare R2 storage and snapshotted daily for 14 days. Agent tokens are stored only as one-way hashes.
What we send to others
Two automatic checks help agents do better work. Both can be switched off per workspace in Settings → AI checks.
| Service | When | What is sent |
|---|---|---|
| OpenAI | After a page is saved (page check) | A screenshot of the page, the text visible on it, and the names and descriptions of its datasets and columns. Raw rows are not sent, but anything shown on the page is in the screenshot. OpenAI's API terms say it does not train on this. |
| TypeSafe | Before a new page is created (request check) | The brief the agent wrote: what you asked for, what it asked back, and the questions the page should answer. |
| When you sign in | Nothing beyond the sign-in. We ask only for your name and email. | |
| Stripe | When you subscribe to a paid plan | Your email address. You give your card and billing address to Stripe directly; we never see or store the card. We keep Stripe's customer and subscription IDs, your plan and its renewal date. |
| Resend | When we email you | Your email address and the email: page titles, dataset and agent names, counts and links, never rows. We email the person who connects an app the day after if nothing was built yet, a workspace's admins when a dataset misses its schedule, and members a weekly summary of the pages they can open. Every email has a link that stops them all. |
| Fly.io, Cloudflare | Always | Hosting, backups and delivery. Cloudflare caches the built code of pages (never their data). On the home, pricing, connect, sign-in, privacy and terms pages, Cloudflare Web Analytics counts visits without cookies. On the ChatGPT page, Quillm also counts visits and button clicks per day and campaign tag: a number, nothing about you. |
| esm.sh | When an agent declares an extra npm package | The package name and version, to download it. Nothing about you or your data. |
| Anyone with a share link | When a page is shared “with the link” | That page and the data it shows, read-only. Admins can turn links off for a whole workspace in Settings → Sharing. |
What we never do
Sell your data, train models on it, or show it to anyone outside your workspace, except the pages you chose to share by link. We look at a workspace's contents only to fix a problem you report, or when a shared page is reported as abuse.
Your data is yours
Settings → Export downloads everything in a workspace: every page's source and revisions, every dataset as CSV, notes, members and the change log. Deleting a workspace in Settings removes it from the database at once and from backups within 14 days. To delete your account, write to [email protected].
Cookies
One cookie keeps you signed in for 30 days. There are no analytics or advertising cookies.