AI processing and consent
With your explicit, versioned Infernal AI consent, selected photos, item details and valuation messages are sent to OpenAI for identification, moderation, market research, estimates, explanations, translations and chat. Research can use web search. Do not submit unnecessary personal information. Consent can be withdrawn in Privacy settings. New work stops and unfinished evaluations are cancelled and their reserved credits released; provider calls already admitted may finish. Withdrawal does not erase existing history. Shared NewTopic AI consent is separate and its scope is disclosed independently.
Community data and notifications
Public comments and replies show your shared display name, username and avatar. Likes, reports, blocks, moderation decisions, translations and activity records support community features. Your private moderation-help conversation is not published. With separate community AI consent, OpenAI processes comment text and relevant public discussion context for safety and usefulness review, translation and moderation help. Withdrawal stops new community AI work and cancels pending work, without automatically deleting approved comments. These records are included in account exports and membership removal. Comment drafts are stored only in the current browser session and cleared on account switching or sign-out. Optional iOS notifications use Apple Push Notification service and a device token tied to the signed-in account; push payloads contain minimal activity identifiers, not comment text.
AI processing records
Valuation research can be processed in the background. The provider retains the response needed to complete and retrieve that work; Infernal AI records its identifier and requests deletion during membership erasure. Community moderation, translation and help requests do not ask the provider to save a retrievable response history. This does not mean that the provider keeps no operational or abuse-prevention records. OpenAI’s applicable retention and data-protection terms govern that processing. Human support and legal requests are not sent as AI prompts.

