Privacy Policy

This policy covers two groups of people:

• Business owners — the people who create an agenthelp account and run an assistant for their business.
• Customers — the people who chat with a business's assistant, call its AI phone line, or book an appointment through it. Customers don't create accounts with us; when they talk to an assistant, they're talking to the business, and we process their information on that business's behalf.

The Service is intended only for businesses located in the United States and their customers in the United States. We don't knowingly offer it to, or collect information from, people in the EU/EEA or the United Kingdom.

From business owners:

• Account details — business name and email address.
• Your password — stored only as a one-way cryptographic hash. We cannot read it, and we never store it in plain text.
• The business information you give your assistant — services, hours, prices, policies, and similar setup details.
• Your business address — when you turn on the AI phone service, so we can register the phone number and emergency (E911) service the law requires.
• Billing — payments are processed by Stripe. Your full card number never touches our servers; we keep only what Stripe gives us to manage your subscription.

From customers of a business:

• Chat messages exchanged with the assistant (every conversation discloses it's an AI).
• Phone-call information when a business uses the AI phone service — the caller's phone number, the number dialed, and basic call details (time and length). See “AI phone calls and call recording” below for how recordings and transcripts are handled.
• A chat identifier stored in your browser — only after you send a message — so that if you return to the same business's chat on the same device within about two weeks, your conversation picks up where it left off. You can clear it anytime with the chat's “New chat” button or by clearing your browser data.
• Booking details — name, phone number, and the appointment requested — plus whether the one-time text-message code used to confirm the phone number was verified.
• How you arrived at the chat — the site that referred you and any campaign tags (“UTM” parameters) in the link you followed — so the business understands which of its links bring in customers.

From everyone, we collect basic technical logs (such as IP address and timestamps) to keep the Service secure and rate-limit abuse.

• To run the Service — answer questions, answer phone calls, book appointments, and capture leads.
• To send service messages — booking confirmations and appointment reminders by email or text (these are transactional, not marketing).
• To prevent fraud and abuse — rate limits, phone verification, and security monitoring.
• To bill subscriptions and provide support.

We do not sell personal information, use it for advertising, or use conversation content to train AI models.

When someone sends the assistant a message or speaks to the AI phone line, that message or speech (plus the business's setup information and the conversation so far) is sent to our AI providers to generate the reply. Anthropic generates the assistant's wording and is not permitted to use it to train its models; for phone calls, Vapi provides the voice platform that turns speech into text and back. They process the information to produce the response, not to build their own products.

If a business turns on the AI phone service, an AI answers its calls. Every call opens by telling the caller it's an AI and that the call may be recorded, before anything else is discussed — so callers know up front and can choose to continue or hang up.

Call audio, any recording, and the transcript are held by our voice provider (Vapi) and the telephone carrier (Twilio); we do not store call recordings or transcripts in our own database. What we keep is limited call information — the caller's number, the number dialed, the call status, and how long it lasted — used to route the call, prevent abuse, and bill the service.

We share data only with the providers that run the Service, and only what each one needs:

• Anthropic — generates the assistant's replies (chat and phone). Contractually prohibited from training on your data.
• Vapi — powers the AI phone line (speech-to-text, text-to-speech, call audio and transcripts).
• Twilio — provides phone numbers and sends text messages (verification codes and appointment reminders).
• Stripe — processes subscription payments.
• Resend — delivers our service emails, such as booking confirmations and reminders.
• Upstash — runs the rate-limiting that protects the Service from abuse (counters and IP addresses, no customer profiles).
• Sentry — collects error reports so we can fix problems (configured to strip personal information from those reports).
• Neon — hosts the database (encrypted at rest).
• Vercel — hosts the application.

Each provider acts on our instructions under a data-protection agreement; none of them are data brokers and none sell your data. Our current list of providers is also published at getagenthelp.com/legal/subprocessors. Beyond these providers, we disclose personal information only if the law requires it or to protect the safety and integrity of the Service.

• Anonymous or abandoned chats — deleted automatically after about 30 days.
• Other conversation transcripts — kept up to about 12 months, and businesses can shorten that window or delete transcripts anytime from their dashboard.
• Bookings and account data — kept while the business's account is active.
• When an account closes, we delete its data within a reasonable period, except what we must keep for legal or accounting reasons (for example, billing records, which tax law requires us to retain for several years).

Every business's data is isolated from every other business's — that separation is enforced at the database level, not just in the application. Data is encrypted both in transit and at rest (AES-256), passwords are stored only as one-way hashes, calendar connections are encrypted, and secret keys never ship to the browser. No system is perfectly secure, but the Service was built with these protections from day one.

• Business owners can edit their business information, delete chat transcripts, and close their account — which deletes its data — from the dashboard, or email us for help with any other request.
• Customers can ask the business they interacted with to delete their information; the business can remove conversations from its dashboard, and can have other records (such as a booking or lead) deleted by contacting us. You can also contact us directly and we'll help.
• Depending on where you live, you may have legal rights to access, correct, delete, or receive a copy of your personal information. Email us and we'll honor the rights that apply to you.

The Service is for businesses and their customers; it is not directed to children under 13, and we don't knowingly collect their information. If you believe a child has given us personal information, email us and we'll delete it.

We may update this policy as the Service evolves. If a change is material, we'll notify business owners (for example, by email or in the dashboard) before it takes effect. The "Last updated" date at the top always tells you when it last changed.

Questions, or a privacy request? Email support@getagenthelp.com. Our Terms of Service explain the agreement that governs the Service itself.