Privacy
Last updated 9 October 2026
Yuddle is in private beta. This page says, in plain words, what we keep, why, where it lives and who helps us run it. We don’t sell your data, show ads, or let anyone train models on it.
What we keep
- On this website: if you join the waitlist, your email address, the page that sent you here and your browser type. If you apply as a design partner, your answers (name, company, team size, the AI tools you use, and your note). Page visits are counted with Vercel Web Analytics, which doesn’t use cookies.
- Your account: your name, email address and profile picture from the way you sign in (Google, GitHub or an email link).
- What you add to Yuddle: notes, files, chat exports, the Google Drive files you pick, the GitHub issues you follow, and Claude Code session digests if you install the plugin (it removes secrets on your computer before anything is sent). Also the items in your Yuddles, who proposed and confirmed each, and their history.
- Ask: your questions and the answers, kept in your own conversations for 90 days unless you delete them first.
- Records about use: which tools an AI you connected called and which items it read (as IDs and counts, never the text), the days you opened a Yuddle, emails we sent, and rate-limit counters.
Who sees it
People see only the Yuddles they belong to, and only the sources they’re allowed to open. Anything you keep private is seen by you alone. An AI you connect (Claude Code, or Claude through its connector) reads as you: exactly what you can see, nothing more. Your Claude inbox and your Ask conversations are yours alone.
How AI is used
A language model reads what you add and proposes changes to your Yuddle; a person on the team confirms each one. Ask answers your questions from what you can see.
- We use Anthropic’s Claude through the Vercel AI Gateway with zero data retention, so the request isn’t stored by the gateway or the model provider.
- If the gateway is unavailable, a request may go to Anthropic’s API directly or to Microsoft Azure OpenAI, under their business terms.
- Your data is never used to train models.
Where it lives, and who helps us
Everything is stored and processed in the United States, by:
- Vercel: hosting, the AI Gateway and website analytics
- MongoDB Atlas: the database
- Google Cloud: our sign-in service
- Anthropic and Microsoft Azure OpenAI: the language models
- Resend: email
- Inngest: background jobs, which carry IDs rather than your content
Connection tokens for Google Drive and GitHub are encrypted. Device and read tokens are stored only as fingerprints.
How long we keep it
During the private beta we keep what you add until you remove it, or until you ask us to delete it. Ask conversations are deleted after 90 days. Waitlist entries stay until you unsubscribe or ask us to remove them. History in a Yuddle (who confirmed what, and when) is kept so the team can always see why something is the way it is.
Your choices
- Remove a source from a Yuddle; what rested only on it is withdrawn.
- Delete an Ask conversation, leave a Yuddle, or disconnect an AI or a device.
- Owners can export a Yuddle as a file.
- Unsubscribe from any email with the link at its foot.
- Email us to see what we hold about you, or to delete your account and what only you added.
Questions, or want something removed? Email hello@yuddle.app.