1. Overview & Commitment
PageClerk ("we", "our", or "us"), operated by Reshapify, LLC (1111B S Governors Ave STE 23005, Dover, DE 19904, United States), is committed to safeguarding the privacy and security of your personal data and the organizational documentation you entrust to our platform. This Privacy Policy explains what information we collect, how it is processed to deliver AI assistant responses, and your rights regarding your data.
Chatting with us on pageclerk.com: the assistant on our own website runs on PageClerk like any customer's. For those conversations we are the controller: we use your messages and any details you leave (such as your name, company, website and phone number) only to answer you and to follow up about PageClerk. Conversations are deleted after 365 days; details you leave are kept until you ask us to delete them at privacy@pageclerk.com.
2. Information We Collect
We collect information in the following categories:
- Account Credentials: Name, email address, company name, and passwords, which are stored only as salted one-way hashes and never in readable form. If you sign in with Google, we receive your name, email address and profile picture from Google, and don't receive your Google password. One-time sign-in codes we email you are stored only as hashes and expire after 10 minutes.
- Website Analytics: On our public pages (not the dashboard, and not customers' chat widgets or hosted chat pages), we count visits with analytics software we host ourselves, which doesn't use cookies. It records the page, referring site, browser, device type, screen size, language and approximate location (country, region and city). Your IP address is used only to work out that location and an anonymous visit ID that changes daily, and isn't stored. Query strings in page addresses are removed before anything is sent.
- Customer Knowledge Base Sources: Content extracted from URLs, XML sitemaps, uploaded PDF, DOCX, TXT, or Markdown files, and custom Q&A pairs.
- Search Index: Numerical representations of your source text, used to find the passages relevant to each question.
- Conversation Transcripts: Messages, queries, assistant replies, latency statistics, cited source tags, and any 👍/👎 ratings visitors give replies, recorded during chat interactions.
- Visitor Lead Inquiries: Contact information (name, email address, phone number) and answers to any extra questions our customer adds to their lead form, submitted voluntarily by visitors through the chat widget.
- Bookings and Handoffs: When a customer connects a booking or support tool, the details a visitor gives to book an appointment or reach a person (such as name, email, chosen time and notes), and the resulting booking or ticket reference.
- Messages from Connected Channels: Messages visitors send through channels a customer connects, such as Freshchat, which are answered and stored like website conversations.
- Conversation Signals: Automated classifications of conversations described in section 3, such as an intent or urgency label, a spam score, or a rating of how well a question was answered.
- Billing Information: Payment card numbers and transaction details processed securely via Stripe. We do not store raw card numbers on our servers.
3. AI Data Processing & Model Training
We do NOT use your confidential documents, uploaded files, or private customer conversation transcripts to train or fine-tune public foundational AI models.
Your content is utilized solely to perform Retrieval-Augmented Generation (RAG) on behalf of your account:
- Your source text is split into passages and indexed for search.
- When a visitor asks a question, we search only the sources of that chatbot.
- The retrieved relevant snippets are passed as contextual references in the API call to generate an accurate, grounded answer citing your sources. Replies are generated by the AI model providers listed in section 4, depending on the answer quality a customer chooses. Customers who connect their own model provider account (OpenAI, Google, xAI, Anthropic, Mistral or DeepSeek) have replies generated by that provider instead.
Automated classification
Conversations are also classified by an automated decision model from TypeSafe, which returns labels and probabilities rather than text. Depending on the features a customer turns on, it is used to:
- filter spam, attempts to manipulate the assistant, and off-topic misuse, which receive a standard reply instead of an AI answer;
- choose which model answers a question (the "Auto" answer quality);
- decide when to offer the lead form or a handoff to a person;
- label leads with intent, urgency and how likely they are to be a real prospect;
- rate how well finished conversations were answered, for the customer's reports;
- check that an action the assistant is about to take, such as making a booking, was requested by the visitor.
These classifications support our customers' own decisions about how to follow up with visitors. They do not produce legal or similarly significant effects for visitors, and a visitor can always reach the business through its other contact routes.
4. Third-Party Service Providers
We work with trusted infrastructure providers to deliver our service:
- Convex: Database, backend hosting and search indexing, in the United States.
- Convex AI Gateway and OpenRouter: Route requests to the AI model providers below.
- Google, xAI and OpenAI: AI models that generate replies. Google also builds the search index from customers' sources.
- Cloudflare: Website hosting, and fetching and rendering the web pages customers ask us to index.
- Stripe: Payment processing, customer billing management, and tax calculation.
- Resend: Transactional emails, including sign-in codes, password resets, email verification, lead and handoff notifications, and the weekly summary customers can turn off.
- TypeSafe: Classifying conversations for spam and abuse filtering, routing, lead tagging, handoff timing, answer grading, and safety checks on actions.
- Customer-enabled integrations (Cal.com, Slack, HubSpot, Freshdesk, Freshchat, webhook endpoints such as Zapier, n8n, Make or HighLevel, model providers chosen by the customer, and customers' own APIs): Only when a customer turns them on. Visitor details and conversation content are sent to the services that customer chooses, to book appointments, send alerts, add contacts to a CRM, open support tickets, answer chats, route leads or look up information. Those services process the data under their own terms and the customer's agreement with them.
5. Data Retention & Deletion
You maintain complete ownership of your data at all times. You can delete individual conversations and leads, sources, or a whole chatbot from your dashboard; deleted data is permanently removed from the database. You may also request full account closure and data deletion by contacting support, and we delete your data within 30 days.
Conversation history is deleted automatically once it is older than the period included in your plan (from 7 to 365 days). Leads are kept until you delete them. Records of automated classifications kept to tune and audit them are deleted after 90 days, and records used to avoid answering a Freshchat message twice after 7 days. Data already sent to an integration a customer connected is held by that service under its own policies.
6. Security Measures
Data is encrypted in transit (TLS) and at rest by our hosting providers. Sessions use secure HTTP-only cookies, and every request is scoped to the signed-in account, so each account's data is kept separate. API keys customers give us for model providers and integrations are additionally encrypted with AES-256-GCM and never shown again. Requests we send to customers' webhooks and APIs are signed, and webhooks we receive from Freshchat are verified against the customer's account key.
7. Your Rights (GDPR & CCPA)
Depending on your jurisdiction, you possess the rights to:
- Access and receive a copy of your stored personal information.
- Export your conversation transcripts and lead data in CSV or JSON format.
- Delete individual conversations and leads, for example when a visitor asks you to erase their data.
- Rectify inaccurate personal details via your account settings.
- Request the complete erasure of your account and all associated data.
8. Contact Us
Customers: our Data Processing Agreement covers how we process your visitors' data on your behalf, and lists our sub-processors. If you have any questions, inquiries, or requests regarding this Privacy Policy or your data, please contact us at: privacy@pageclerk.com