We never train on your data
Your training content, conversations and leads are used to answer your visitors, and for nothing else. We do not train or fine-tune AI models on them.
Security
An AI product should hold itself to a higher bar on data. Here is ours.
Your training content, conversations and leads are used to answer your visitors, and for nothing else. We do not train or fine-tune AI models on them.
Questions and passages sent to OpenAI to write an answer go through its API, and OpenAI does not train its models on API data.
Download everything stored for your account, or delete the account and all of it, under Settings → Your data. No lock-in.
The safeguards built into every account and every chatbot.
Every page, the chatbot and the dashboard are served only over HTTPS, with HSTS so browsers never fall back to an unencrypted connection.
Every website has its own chatbot, training content, conversations and leads. Only the account that owns a website can see its data.
Passwords are stored hashed, never in plain text. Email addresses are verified with a one-time code, and sign-in attempts are rate limited per network and per account.
A strict Content Security Policy, clickjacking protection and locked-down browser permissions are sent on every page.
The public chatbot limits messages per visitor, per network and per chatbot, and only opens from your own website and the domains you allow.
We use a short list of providers to run the service, and publish every one of them with what it does.
Exactly what we store, where it goes, and how long it stays.
Conversations are stored so you can read them and improve answers. You can set a retention period in the chatbot settings, after which old transcripts are deleted automatically.
Names, emails and phone numbers that visitors choose to share are stored with their conversation and shown only to the website owner.
When you upload a file, we keep the text extracted from it to train your chatbot, not a public copy of the file. Remove a source at any time and it stops being used.
Deleting a website permanently removes its chatbot, training content and conversations. Leads it collected stay in your account until you delete them, or delete the whole account under Settings → Your data. Need help? Write to support@patchwhistle.com.
To answer a question, we send the question and the most relevant passages from your content to OpenAI, which writes the answer.
The chatbot is told to use only the passages we send, and we check answers before visitors see them: a price, phone number, email or link that is not in your content is removed.
Each chatbot stops at its plan’s monthly message limit, and we email you when it is reached.
Browsers can only open your chatbot from your own website (with or without www), any extra domains you allow in its settings, and its shareable chat page if you turn that page on.
Training crawls and Monitoring checks only visit publicly reachable pages on the domain you entered. Private, login-protected, local and off-domain addresses are skipped.
Payments are handled by Dodo Payments; card details never reach us. Plan changes are applied from verified payment webhooks, not from the browser.
Everything you need for your own security or privacy review.
Every third-party service that processes data on our behalf, and what it does.
View listWhat we collect, why, how long we keep it, and your rights.
Read noticeThe agreement that covers your use of Patchwhistle.
Read termsFilling in a vendor questionnaire, or found a vulnerability? Write to us and we will reply quickly.
support@patchwhistle.comPaste your website address, test the answers, and put it live with one line of code. Free for 14 days, no card.