Website inquiry content — 30 days
The website setting makes the website’s stored contact, event and message content eligible for redaction 30 days after the inquiry is accepted. Both the original website payload and the delivery-job payloads are covered. Redaction removes that content from the active application records; it is not deletion of every record associated with the inquiry.
Cleanup runs on a schedule and in batches. Processing safeguards can postpone a record that is actively being handled. The period is measured from acceptance, not from the last team reply. The website uses this 30-day eligibility setting. Cleanup is scheduled daily; this is not a promise of deletion at an exact instant.
Delivery receipts and duplicate prevention
After content is redacted, the system can retain inquiry and job identifiers, timestamps, delivery status, attempt counts, destination record references and protected comparison values. These records support troubleshooting, accountability and preventing an old submission from creating a second lead.
They are not described as anonymous: a destination reference or protected identifier can still be associated with a record elsewhere. The current implementation does not give all such metadata a fixed automatic expiry. Its continuing necessity and retention must be reviewed separately; the 30-day payload setting is not a claim that every identifier expires after 30 days.
Security and operational records
The application cleanup routine removes rate-limit records older than one day and its audit events older than 90 days. Those periods describe eligibility during scheduled cleanup, not a guarantee that each entry disappears at that exact age. Hosting, infrastructure and provider logs have their own retention settings.
Operational monitoring is designed to report aggregate counts and safe condition names, not customer messages. Restricting what goes into logs reduces exposure but does not make every technical record non-personal.
Copper, Sheets, email and bookings
The website’s cleanup does not delete a Copper lead, a Google Sheets row, a team email, an existing workflow’s output or a signed agreement. Those records support inquiry follow-up and, where applicable, booking administration, accounting, dispute handling and legal obligations.
Their appropriate retention depends on the record’s purpose, whether the inquiry remains active or became a booking, applicable requirements and the system holding it. There is no single period in this notice that applies to every business record. Contact the team if you would like to ask about a particular record or request its deletion.
Browser storage and analytics
The inquiry security cookie lasts up to 20 minutes. Session-storage records normally last for the tab’s session, subject to browser restore behavior. Marv’s saved conversation also expires after 24 hours without an update. Clearing local site data does not delete server-side business records. See Cookies & tracking.
Google Analytics and advertising tags are disabled on this website at launch. The retention of historical records in existing Google accounts depends on those accounts and is not changed by this notice.
Backups, legal obligations and requests
A change in an active database may remain in historical backups until those backups expire under the provider’s arrangements. The website cleanup routine is not an immediate purge of provider backups. Separate preservation obligations may apply to particular business records; they must be handled deliberately rather than assumed to be covered by a general retention setting.
To ask about a record or request correction or deletion, email inquiry@marvimon.com with “Privacy request” in the subject. We may need to verify the request and consider applicable exceptions. Read Privacy policy for the request process. Please do not assume that clearing browser data, withdrawing an inquiry or removing its website copy deletes information everywhere.
Marv chat and usage records
Before a staff follow-up is submitted, Marv’s conversation, unsent chat draft and panel preference are saved in the tab’s session storage, not in the website inquiry database. The complete current conversation is retained locally, up to 201 messages and 100,000 characters; new questions stop at that safety limit rather than removing older messages. A restored older conversation may be incomplete and is labeled accordingly. The saved conversation expires after 24 hours without an update, or when the tab’s session storage is cleared; browser restore features may preserve tab sessions. Starting a new conversation removes the saved chat. Handoff contact fields and submission state are not persisted or replayed. OpenAI separately processes the question and a limited conversation context when AI chat is enabled; provider retention is described at Privacy policy. Clearing or resetting the chat does not delete information already processed by a provider.
Protected visitor identifiers and token/cost records support abuse prevention and the shared spending limit. They do not contain questions, replies or handoff contact fields. Their current implementation has no fixed automatic deletion period. The website inquiry-content cleanup does not erase this usage history or provider records.
A staff follow-up becomes a separate inquiry only after the visitor reviews and submits it. When transcript sharing is enabled, the visitor reviews the available conversation before sharing it in separate Copper notes. The encrypted website delivery copy, including that transcript, follows the inquiry-content setting; Copper, Google Sheets and any subsequent correspondence have separate retention rules. The live website creates real inquiry records only after an explicit submission.