Privacy Policy
Last updated October 14, 2026
This policy explains how Davos Dev LLC (“tufu”, “we”, “us”) handles personal information in connection with tufu.ai and the tufu service (the “Service”).
1. Two roles
- Our customers and our website. For the people who sign up for tufu and use it (“Users”) and for visitors to tufu.ai, we decide how personal information is used. We are the controller, and this policy applies in full.
- Our customers' visitors. When someone chats with a bot on a customer's website, or wherever else the customer connects it (a “Visitor”), we process that conversation on behalf of the customer, who decides what the bot does. The customer is the controller and we are its processor, under our Data Processing Addendum. If you chatted with a bot on another company's website, that company's privacy notice applies, and you should send requests about your data to them. We will help them respond.
2. What we collect
From Users
- Account details: your email address, name, profile picture and time zone, and your sign-in details. Passwords are stored only as secure hashes. We send a link to confirm your email address before you can sign in. If you sign in with Google, we receive your name, email address and profile picture from Google.
- Team and settings: your teams and roles, bot settings, notification preferences, and the Slack workspace and channels you connect, and the Discord server and channels you connect (their IDs and names). To route Slack replies to the right person, we match your Slack profile email to your tufu account.
- Your content: the knowledge you add (files, website pages, repositories and other connected sources), and the replies, notes and actions your team takes in conversations. A reply you send to a Discord thread is posted there with your first name.
- Email replies: when you reply to a notification email, we receive the message and its sender address to post the reply in the conversation.
- Email delivery: whether the emails we send you were delivered, bounced or marked as spam. If an address bounces or reports us as spam, we pause optional and notification email to it until you ask us to try again.
- Billing details: payments (subscriptions and answer packs) are handled by Stripe, which receives your card details directly as its own controller under its privacy policy. We keep your plan, subscription status, purchases and invoices, never your full card number.
From Visitors (on behalf of our customers)
- Conversations: the messages a Visitor sends, the bot's answers and the replies from the customer's team.
- Contact details: an email address, if the Visitor gives one so the team can reply after they leave.
- Support emails: if a customer forwards its support address to tufu, the emails its customers send: the sender's name and address, the subject, the text, the names of any attachments (we don't keep the files), message identifiers used to keep the conversation threaded, and what the receiving server reported about the sender's authenticity (SPF, DKIM, DMARC). We also record whether the replies tufu sends were delivered, bounced or marked as spam.
- Technical details: the IP address and a random identifier stored in the Visitor's browser (so the conversation continues when the page reloads). Both are also used, for a few hours at most, to count questions against our usage limits and prevent abuse.
From Discord (on behalf of our customers)
A customer can add tufu's Discord app to its Discord server, so that its bot answers members' questions there. As with Visitors, the customer is the controller of these conversations and we are its processor (section 1).
- What Discord sends us: the messages posted in the channels and threads the app can see in that server, each with its author's Discord user ID and username. We don't receive direct messages, and we don't read attachments, profile pictures, display names or member lists.
- What we keep: only the messages meant for tufu. That means messages that mention the app, uses
of its
/askand/humancommands, clicks on its buttons (“Talk to a human” and the helpful or not helpful votes), and every message posted in a thread the app started or answered in, including while that thread waits for a person. If the customer's plan doesn't include Discord, we don't keep the message, but we email the customer's admins the asker's username. For these we store the sender's Discord user ID and username; the server, channel, thread and message IDs; the message text; the bot's answers and the team's replies; and who voted on which answer. Every other message is dropped as it arrives, without being stored or logged. - How we use and share it: to answer the question and continue the conversation, to hand it to the customer's team and notify them (by email, or in Slack if the team connected it), and to show the team which questions its knowledge didn't answer. To write an answer, we send the conversation to the AI provider and the question to our search provider (section 3). Discord user IDs are also used, for a few hours at most, to count questions and requests for a person against our usage limits and to prevent abuse. We don't use Discord messages to train AI models, we don't sell them, and we share them with no one else except as section 5 describes.
- What we post to Discord: the bot's answers and their sources, the team's replies (signed with the replying person's first name and the team's name), short status lines, and a thread title and an “@username asked” line made from the question. Everyone who can see that channel can read them.
- How long we keep it, and deleting it: these conversations belong to the customer's account. We keep them as long as the customer keeps them in tufu, including after it removes the app from its server, and delete them when the customer deletes its team (section 7). If you are a member of a server that uses tufu and want your messages deleted, ask that server's team, or write to us at with your Discord username, your Discord user ID if you can, and the server's name. We will delete the messages and votes we can find from you and let the server's team know.
Automatically
- Logs and security: IP addresses, browser type, request and error logs, used to run and secure the Service. When something breaks, our error monitoring (Sentry) receives details of the error, such as the page and the code involved and which account and team it happened in (by internal identifiers). It is set up not to collect email addresses or IP addresses. Sign-up is protected by Cloudflare Turnstile, which checks that a person, not a script, is signing up.
- Product analytics: on tufu.ai and inside the tufu app, we may record which pages are viewed and which features are used, without cookies (see section 6). We never run our analytics in the chat or Ask windows embedded on customers' websites.
3. How we use it
- to provide the Service: answering Visitors, handing conversations to the right people, and sending notifications by email or Slack;
- to generate AI answers, summaries and suggested replies, by sending the relevant conversation and knowledge to our AI provider, and to make knowledge searchable, by sending it to our search (embeddings) provider;
- to answer support emails a customer forwards to tufu, and to send the replies in that customer's name;
- to secure the Service, prevent abuse and spam, and enforce our Terms;
- to communicate with Users about their account, the Service and changes to it;
- to understand how the Service is used and improve it; and
- to meet legal obligations.
No training on your data. We do not use Users' content or Visitors' conversations to train AI models. Our AI providers (Anthropic for answers, Voyage AI for search) process this data under terms and settings that do not allow them to train on it; they may keep it for a limited time to detect abuse. If a team adds its own AI provider key, that bot's requests go through the team's own account, under the team's agreement with that provider. We do not sell personal information or share it for cross-context behavioral advertising.
4. Legal bases (EEA, UK and Switzerland)
- Contract: to provide the Service to Users under our Terms.
- Legitimate interests: to secure and improve the Service, prevent abuse, and understand usage in a way that doesn't identify you through cookies, where those interests aren't overridden by your rights.
- Legal obligation: for tax, accounting and lawful requests.
- Consent: where we ask for it. You can withdraw consent at any time.
For Visitors' data, the customer determines the legal basis.
5. Who we share it with
We share personal information only with providers that help us run the Service (our subprocessors), with the integrations a team chooses to connect (such as Slack or Discord), when the law requires it, to protect rights and safety, or as part of a merger or acquisition (in which case this policy keeps applying to your information).
6. Cookies and similar technologies
- Necessary: cookies that keep you signed in and protect forms (session and CSRF), and that remember your theme and language. The chat widget stores a random identifier in the Visitor's browser so a conversation survives a page reload. These are needed for features the person asked for.
- Analytics: our product analytics run without cookies or device storage, and honor your browser's Do Not Track setting. They are not loaded in the embedded chat or Ask windows.
- No advertising cookies, and we don't let third parties set cookies through our site for their own purposes.
7. How long we keep it
- Account and team data while the account or team exists. When a team is deleted, its content, including Visitors' conversations, is deleted from our live systems within 30 days and from backups as they expire, within 90 days.
- Visitors' conversations as long as the customer keeps them in tufu, unless deleted sooner.
- Logs for up to 90 days, unless needed longer to investigate abuse or security incidents.
- Billing records as long as tax and accounting law requires (typically up to seven years).
8. Your rights
Depending on where you live (for example under the GDPR, UK GDPR or California law), you may have the right to access, correct, delete, or receive a copy of your personal information, to object to or restrict some processing, and to withdraw consent. You can update most account details on your profile page. For anything else, email . We respond within one month, and we won't treat you differently for exercising your rights. You can also complain to your local data protection authority. Visitors should contact the company whose bot they used (on Discord, that server's team). We will pass along requests we receive.
9. Security
We encrypt data in transit, limit access to people who need it, offer two-factor authentication, and rely on providers that encrypt data at rest. No system is perfectly secure. If a breach affects your personal information, we will notify you and the relevant authorities as the law requires, without undue delay.
10. International transfers
We and our providers process data in the United States and other countries, which may have different data protection laws than yours. Where the law requires it, we protect transfers with safeguards such as the European Commission's Standard Contractual Clauses and the UK Addendum.
11. Children
The Service is for businesses and is not directed to children. Users must be at least 18. We don't knowingly collect personal information from children under 16. If you believe a child has given us personal information, contact us and we will delete it.
12. Changes
We will post changes here with a new date. For material changes we will tell Users by email or in the Service before they take effect.
13. Contact
Davos Dev LLC, privacy questions and requests: .