Oernoe AI is scheduled for release on November 1, 2026. This guide explains the published data policy and how to prepare a safer prompt. It does not promise that an unreleased control is available or that a successful response has been verified.
A conversation may be stored
The Oernoe AI Privacy and Data Policy says Oernoe may collect prompts, files, conversations, settings, and technical information. If memory features are enabled, selected information may also be kept to personalize future responses. This is different from a promise that every prompt disappears when a response ends.
Deleting a memory may not delete its source conversation. Account deletion, individual conversation deletion, and memory controls can have different scopes. Some backup, security, or legally required records may remain for a period described by the applicable policy.
Reduce the information in a prompt
Write the task first, then add only the context needed to answer it. Replace identifiers with neutral labels. For example, ask “Summarize the cancellation clause in this agreement” and provide the relevant clause with names, signatures, account numbers, and unrelated terms removed.
- Replace a customer’s name with “Customer A” when the identity does not matter.
- Remove passwords, API keys, recovery codes, and private keys entirely.
- Use a small representative code example instead of an entire production repository.
- Do not upload health, financial, or identity records unless the feature specifically supports that use and you have reviewed its notice.
These are actions you take before sending. Do not assume that automatic masking will remove every sensitive detail.
Read the training and connected-source notices
The AI policy ties training use to the choices and notices shown in the product. Review those notices rather than relying on a network-wide slogan. If a separate research or training opt-in is offered, the policy says it should be voluntary and clearly labeled.
A feature that uses Search or a connected source may send a query and limited context to obtain results. A connected provider has its own terms and practices. Before connecting a service, check what information the feature needs and whether you have permission to send it.
Verify a response before using it
Ask for sources and open them. Check that each source supports the claim attributed to it. A citation, a model name, or a confident answer is not proof of accuracy. Test generated code in an appropriate environment and get qualified advice before relying on medical, legal, or financial conclusions.
Request help or deletion
Contact privacy@oernoe.com and identify the product and the data or action concerned. Do not resend a private conversation just to ask how deletion works. The privacy request guide explains how to scope a request. Read the AI terms and usage and safety policy before using the service.