Effective date: July 1, 2026 · Last updated: September 21, 2026
CapeFear.ai ("CapeFear.ai," "we," "us," or "our") provides managed agentic‑AI services to small and mid‑sized businesses. This Privacy Policy explains what information we collect, how we use and share it, and the terms that apply to our text‑message (SMS) communications. By using our website or opting in to our messaging, you agree to this policy.
We (and businesses that use our platform, such as participating dental and professional‑service practices) may send text messages to individuals who have opted in. The program includes:
Mobile information is never shared for marketing. No mobile information (including phone numbers collected for text messaging and consent) will be shared with or sold to third parties or affiliates for marketing or promotional purposes. Mobile opt‑in data and consent are not shared with any third parties. We may share information with subcontractors that support our services (for example, message‑delivery providers) solely to deliver the messages you requested.
We do not sell your personal information. We share information only:
As stated above, text‑messaging opt‑in data and consent are excluded from all sharing except as strictly necessary to deliver the messages you requested.
Where a customer asks us to automate work in a system they already use — for example accounting, practice‑management, project‑management, email, or calendar systems — an authorized user of that customer connects the system to us ("Connected Accounts"). This is always an affirmative act by the customer; nothing connects by default.
Google account data. Where a customer connects a Google account, we request only calendar permissions: read‑only access to the calendar, and permission to create and update appointment events, together with the basic sign‑in identity Google provides. Our use of the data we receive is limited to the scheduling workflow the customer asked us to run: we read when the calendar is busy so the assistant only offers appointment times the customer is actually free for; where a customer has told us to book only inside blocks with a particular title, we look up those blocks and keep only their times; we create the appointment events the assistant books on the customer's own calendar, and we read and update those appointments (for example their attendees) when a visit is rescheduled. We do not request access to Gmail, Google Drive, or Google Contacts. We do not sell Google user data, use it for advertising, or use it to train, evaluate, or improve any machine‑learning or artificial‑intelligence model, whether ours or a third party's. CapeFear.ai's use and transfer of information received from Google APIs to any other app will adhere to the Google API Services User Data Policy, including the Limited Use requirements.
How artificial intelligence is used with Connected Account data. Our assistants are powered by Anthropic Claude models served through Amazon Bedrock, a service of Amazon Web Services that we access through our own AWS account; an Amazon Titan model on the same service indexes text for search. The calendar‑derived information a model receives is limited to the open appointment times the assistant offers a caller and, when a caller asks to cancel or change an existing appointment and more than one matches their name, the titles and times of those matching appointments so the assistant can ask which one is meant; other event details are handled by our own software and are not sent to a model. Under Amazon Bedrock's data‑protection terms our inputs and outputs are not used to train or improve models and are not shared with the model providers, and model‑invocation logging is disabled on our account, so the service does not retain prompts or outputs. On phone calls, the caller's speech is transcribed by Deepgram, which receives the audio of the call and no Google account data, and the assistant's replies are voiced by Cartesia, which receives the words the assistant speaks, including any appointment times it offers, and no other Google account data. Deepgram offers a per‑request opt‑out from its model‑improvement program and Cartesia offers an opt‑out from model training under its terms; we use both, and we do not permit either provider to train on our content once those opt‑outs are in effect. A small open‑source voice‑activity model runs inside our own voice‑agent process, which is deployed on LiveKit Cloud, to detect when a caller is speaking; the model is bundled with our software, processes audio in that process, and never sends anything to a model provider. We do not use Connected Account data to train or improve any model.
We use administrative, technical, and physical safeguards appropriate to the sensitivity of the data, and encrypt sensitive data in transit and at rest. We retain information only as long as needed to provide the service and to meet legal obligations, then delete or de‑identify it.
You may request access to, correction of, or deletion of your personal information, and you may opt out of text messages at any time (reply STOP). Contact hello@capefear.ai to exercise these rights.
Our services are for businesses and are not directed to children under 13, and we do not knowingly collect their information.
We may update this policy from time to time; the "last updated" date above reflects the latest version, and material changes will be posted on this page.
CapeFear.ai · Coastal North Carolina · hello@capefear.ai