Privacy Policy

• Account information: your mobile phone number (your account identifier), an optional display name, your time zone, and your messaging channel preference.
• The content you send: the messages, photos, and voice notes you choose to send, and the documents (such as bills, lab results, or insurance papers) you share.
• Health information you share in conversation so your agent can remember it for you — for example, symptoms, conditions, medications, allergies, vaccines, and the providers you see. This is also covered by our Consumer Health Data Privacy Policy.
• A coarse demographic baseline you ask us to remember: your general home area (city and state), your year of birth, and — only when relevant to a screening — your sex. We deliberately keep these coarse and do NOT store your precise street address, ZIP code, or full date of birth, and we do not track your device location.
• Care logistics you ask us to handle: reminders, referrals, appointments, and (if you connect it) limited Google Calendar data — see below.
• Billing information if you subscribe: handled by our payment processor, Stripe; we store a customer identifier, not your full card number.
• Limited technical and usage data: counts and events about how the Service is used (never the content of your messages), and consent records (including the time, version, and where you gave it — on our website, or by your reply in the text thread; website consents also record the IP address and device/browser used).

• To run the conversation and remember what you tell us, so you don’t have to repeat yourself.
• To answer your questions with educational context, using AI models that process your messages to generate replies.
• To use your saved home area to help find care near you, and your year of birth and sex to surface age- and sex-appropriate preventive-screening information — never for advertising.
• To take actions you ask for — finding providers, attempting bookings, calendar events, reminders, and emailing you documents we generate.
• To operate, secure, troubleshoot, and improve the Service, prevent abuse, and comply with law. A small number of authorized team members may review conversations when needed to support, troubleshoot, or secure the Service, under access controls, confidentiality obligations, and access logging.
• Reviewing conversations to improve the product happens only if you separately opt in. We ask you directly, saying no changes nothing about the Service, and you can change your mind at any time by telling your agent.

• Cloud hosting, storage, and database: Amazon Web Services (AI processing and file storage), Supabase (our database), and Railway (application hosting).
• AI model processing: Amazon Web Services (Amazon Bedrock) and Anthropic, which run the models that generate replies and that, only if you separately opt in, help our team review conversations to improve the Service. Both process your messages under commercial terms that prohibit using them to train their models.
• Messaging delivery: our SMS carrier (Twilio) and our iMessage relay (Sendblue), which deliver texts between you and your agent.
• Payments: Stripe, which processes subscription billing.
• Calendar and email (only if you connect them): Google, as described below.
• Web search for current information: a third-party search provider (Exa), which receives only a generalized query with your name and identifying details removed — not your full messages — and an optional re-ranking provider (Voyage) that re-orders the public search results for relevance and receives the generalized query and those results, not your identity.
• Finding and comparing care: when you ask Kite to find or compare nearby doctors, we send a provider’s name and address (never your personal information) to Google’s Places service to retrieve that office’s public ratings and reviews.
• Reference images: when you ask to see what a condition looks like, we send only the condition name (never your personal information) to Wikimedia Commons to fetch a public medical illustration.
• Voice-note transcription and outbound email of your own documents: Amazon Web Services (Transcribe and SES).
• Product analytics and error monitoring: providers that receive event counts and diagnostics only, never the content of your messages or your health data.
• Appointment booking automation (only when you ask us to book): a browser-automation provider (Browserbase) and a computer-use AI model (from Anthropic or OpenAI) used to submit the booking details you authorize (such as your name and phone number, plus your date of birth or insurance if the office requires them) to a provider’s scheduling page.

• Your distilled health memory and records (the saved facts, symptom and measurement logs, photos, profiles, and reminders that ARE the service) are kept for the life of your account, so your agent can remember them — until you delete them or close your account.
• Raw message transcripts age out on a rolling basis (currently about 18 months); operational logs are kept for about 90 days.
• Short-lived caches (such as general web-search and drug-information lookups, which are not tied to your identity) expire on their own within days.
• Billing records are retained as required for tax, accounting, and legal purposes.

Depending on where you live, you may have rights to know about and access the personal information we hold about you, to obtain a copy (portability), to correct it, to delete it, to withdraw consent, and to opt out of any sale, sharing for targeted advertising, or certain profiling. We do not sell or share your information for targeted advertising or use it for such profiling, so there is nothing to opt out of — but you may still exercise your other rights. California residents have the right to limit the use of sensitive personal information (we already limit health data to providing the Service) and, under “Shine the Light,” to ask about disclosures for third-party direct marketing — we make none.

To exercise any right, text your agent or email privacy@grounding.dev — two reliable methods to reach us. We will verify your request (typically by confirming control of your account phone number) and respond within the time the applicable law requires (generally 45 days, extendable once when reasonably necessary). An authorized agent may submit a request with proof of authorization.

Appeals. If we deny your request, you may appeal by emailing privacy@grounding.dev with “Appeal” in the subject line. We will respond in writing within 60 days. If we deny your appeal, Texas residents may submit a complaint to the Texas Attorney General at texasattorneygeneral.gov, and residents of other states may contact their state Attorney General.

The categories of personal information we collect are described in “Information we collect” above. This includes sensitive personal information — your health information and your coarse demographic baseline. We collect it from you directly, use it only for the business purposes of providing and supporting the Service described above, retain it for the periods described in “How long we keep your information,” and do not sell or share it, and do not use or disclose it to infer characteristics about you.

Because we use sensitive personal information only for those permitted purposes, California’s right to limit its use does not change our practices — but you may still contact us to make any California privacy request, including a request to know, access, correct, delete, or limit, at privacy@grounding.dev, and we will not discriminate against you for exercising your rights. Under California’s “Shine the Light” law, we do not disclose personal information to third parties for their own direct marketing.