Privacy Policy

1. Scope and operator

This policy covers LilM, its related support, and these documentation pages. AndWhatNot Studio operates the application and is responsible for its processing described here. Contact: admin@andwhatnotstudio.com. Discord and server administrators have separate responsibilities for their platforms, communities, and copies of content.

2. Information processed and its sources

Information comes from Discord messages, server configuration, the stored knowledgebase, OpenAI, and approved external lookup sources.

• The triggering message, its author’s display name and Discord identifier, server/channel/message identifiers, timestamps, and attached image URLs.
• Up to 50 preceding messages in the channel or thread, limited by the channel’s text budget. Context includes display names and message text; it can include other members’ messages.
• Selected memory facts, preferences, titles, tags, and knowledgebase entries; generated replies in channels configured to save them; and request, error, and database diagnostic information.
• Search queries and retrieved information from approved roleplay wikis where web lookup is enabled. Voice generation is currently disabled in LilM’s configured channels.

3. How information is used

We use this information to understand the conversation, generate replies, recall relevant stored facts, consult approved references, and maintain the bot. LilM can respond without being directly mentioned, depending on channel configuration. Some channels respond to every eligible message; others respond less frequently.

We do not sell personal information, disclose Discord data to advertisers or data brokers, or use Discord message content to train our own AI models.

4. Recipients and service providers

• Discord and channel participants: messages, replies, reactions, and relevant context. Stored memories can be incorporated into future replies to other users in configured channels.
• OpenAI: relevant conversation and image content, display names, the triggering user’s identifier where included in instructions, prompts, retrieved memories, and tool results needed for response generation. OpenAI handles this under its applicable API data controls.
• Approved web sources: in channels with lookup enabled, derived queries can be sent to OpenAI’s search service and directly to the approved NoPixel and ONX Fandom wikis. These requests may include names or details from the conversation. Full channel transcripts are not the request format for direct wiki lookup.

Authorized operators and hosting or infrastructure providers may access information needed to run, secure, and support the application. Information may also be disclosed where required by law or necessary to investigate abuse and protect rights. External services apply their own terms and data controls.

5. Storage, retention, and deletion

LilM does not keep a general local transcript of raw channel history. Routine request-file logging is disabled in its current configuration. Its local memory and reply records are model-generated outputs: selected memories are saved with a category, title, content, and tags, and generated reply text is saved when the channel enables that feature. These save operations do not attach a separate Discord user ID field to each entry.

Generated outputs can summarize, quote, or infer information from the conversation. They may contain names, preferences, or other information relating to identifiable users, and can be inaccurate. Such information remains covered by this policy even though the model generated the wording. General knowledge or fictional lore that does not relate to an identifiable person is not necessarily personal information.

These memory and reply records do not have an automatic expiry in the current bot and remain until manually removed. The current database diagnostics also write query parameters to operational logs, which can include memory text and saved bot replies. Removing a database entry does not automatically remove such log copies. OpenAI’s response storage and monitoring retention are separate and depend on the endpoint and applicable account settings; Discord retains posted messages under its own controls.

Support correspondence and operational records may be retained to resolve requests, investigate failures or abuse, and meet legal obligations. We assess deletion requests against these purposes and any applicable retention requirements. Existing backups and provider-held records may require separate deletion or expiry; we do not promise immediate removal of copies outside our control.

6. Legal bases

Where European or UK data-protection law applies, we rely on performance of a contract for features a user requests; legitimate interests in operating the disclosed service, protecting the community, and maintaining reliability, balanced against affected people’s rights; legal obligations where applicable; and consent where specifically requested and required.

Processing of messages written by other members is not justified merely because another person submits a command. Server administrators must provide appropriate notice and configure the application on a valid basis. You may contact us to object to processing or ask about the basis that applies to your information.

7. Your choices and rights

In LilM’s current configuration, starting a message with ~ prevents that message from triggering a reply and excludes its text from subsequent history assembled by LilM. This does not erase copies already processed or information quoted in other messages. Merely avoiding a mention of the bot does not prevent contextual processing.

You can ask a server administrator which channels are configured, request restricted access or removal of the bot, and request correction or deletion of a stored memory or saved reply. Because memories are stored as facts and text rather than always under an account identifier, include the relevant fact or message reference as well as your user ID.

Depending on your location and the circumstances, you may request access, correction, deletion, restriction, or portability, object to processing, withdraw consent where processing relies on it, or complain to a data-protection authority. Send requests to admin@andwhatnotstudio.com, identifying the bot, your Discord user ID, and the relevant message, channel, clip, or stored fact. We may need reasonable verification; do not send passwords or account tokens.

8. Security and international processing

We use restricted service credentials and access controls to protect the service. No online system can guarantee complete security. Discord, service providers, and infrastructure may process information outside your country. Where data-protection law requires safeguards for international transfers, the relevant processing must use a lawful transfer mechanism. Contact us for information about the arrangements applicable to your data.

9. Documentation website

These documentation pages do not set tracking cookies, load analytics scripts, or require a login. The web server may log technical request information such as IP address, requested page, browser information, and time for delivery, security, and troubleshooting. Following an external link brings you under that destination’s own policies.

10. Children

The application is not directed to children under 13 or anyone below Discord’s minimum age in their country. Contact us if you believe a child’s information has been processed inappropriately.

11. Updates and contact

We may update this policy when features, data practices, or applicable requirements change. The date above identifies this version. Material changes will be communicated through an appropriate service channel where practical. Questions, requests, and complaints: admin@andwhatnotstudio.com.