Privacy Policy

This page documents how the thefire.lol website, and any AI-mediated reference tool deployed on it, handle user data. The policy applies to the static website content, the interactive quizzes, the planned AI reference tool, and any other interactive feature on the site.

The policy is shorter than most because the site collects less than most. Where the site does collect data, the collection is named here and the retention bounded.

Static website browsing

When you browse the static dossier pages, episode pages, essays, bibliography, and other published content on thefire.lol:

• The site uses Cloudflare Web Analytics — a privacy-preserving analytics service that does not use cookies, does not fingerprint browsers, and does not build cross-site user profiles. Aggregate page-view counts and referrer information are visible to the site operator; individual visitor identity is not
• No Google Analytics, no Facebook Pixel, no Adobe Analytics, no advertising trackers
• No cookies are set by the site for tracking purposes; first-party functional cookies may be set if you interact with features that require state (the quiz answers in memory during a session, for example)
• The site is hosted on GitLab Pages; standard server logs (IP address, request URL, user agent) are retained by the hosting provider under their own policy and not accessed by the site operator except for security incident response

This is the lightest-footprint configuration we can run while still understanding what content readers are using.

Email subscription

If you subscribe to the email list (via the embedded Kit / ConvertKit form on most pages):

• Kit (ConvertKit) is the email service provider; your email address is held by Kit under their privacy policy
• The subscription captures your email address and your subscription timestamp; no other personal data is required
• Subscription is double-opt-in; you receive an email asking you to confirm, and only confirmed subscriptions receive the newsletter
• The subscriber list is not sold, not shared with third parties, and is held for the duration of your subscription plus a reasonable cleanup period after unsubscription
• Unsubscribe at any time using the link in any newsletter email; unsubscribed addresses are removed from the active list within Kit’s standard processing window

The subscription serves the newsletter; it is not a source of personal-data monetization for the project.

Quiz interactions

When you take an interactive quiz on the site:

• Quiz answers are held in your browser’s memory during the session and are not transmitted to a server unless the quiz explicitly says so
• The quiz scoring is computed client-side in JavaScript; no server-side record of your individual answers is created
• Aggregate quiz-completion counts may be visible to the site operator via Cloudflare Web Analytics (e.g., “how many people finished the Fake-Ancient-Text quiz”) without tying completion to any individual visitor
• Closing the tab discards your quiz state; the project does not retain your answers across sessions

AI reference tool — if and when deployed

The AI-mediated reference tool described in the AI Reference Tool Design Spec has additional privacy considerations specific to a query-response interface.

Data collected per query

When you submit a query to the reference tool:

• Your query text is transmitted to the project’s server-side infrastructure for processing
• A session identifier is generated to maintain conversation context across multi-turn exchanges within a session; the identifier is not tied to your real-world identity unless you have signed in (sign-in is not required for query use)
• An IP address is visible to the infrastructure as part of standard HTTP request processing
• A timestamp is recorded for the query
• The system’s response is generated and returned to you, with citation links back to the published dossier sections the response drew from

How long it is retained

• Individual query text and response pairs are retained in encrypted operational storage for a maximum of 90 days for service-quality, refusal-pattern monitoring, and refusal-pattern improvement
• After 90 days, individual query-response pairs are deleted from operational storage
• Session identifiers and IP addresses are retained for the same 90-day window and deleted on the same schedule
• Aggregate patterns — anonymized statistics like “this dossier section was cited in N responses last month,” “this category of refusal triggered M times last month,” “the average session length is X turns” — may be retained indefinitely for project research and reporting, with no link back to individual users

What we do with retained data

During the 90-day retention window, the project operator may review query-response pairs for the following limited purposes:

• Refusal-pattern quality assurance — verifying that the tool’s refusal patterns are firing where they should and not firing where they should not
• Safety review — identifying patterns of misuse (attempted operative-instruction extraction, targeted-individual queries, etc.) so the refusal patterns can be improved
• Citation accuracy review — verifying that the responses’ citation links point back to dossier sections that actually contain the claimed material
• Bug investigation — when a user reports a problem with the tool, the operator may review the affected session to diagnose

Retained data is not used for:

• Selling, sharing, or licensing to third parties (the corpus license is non-commercial; query data follows the same non-commercial discipline)
• Training new AI models, whether internal to the project or external
• Building user profiles or attempting to identify individual users
• Marketing or commercial messaging of any kind
• Disclosure to law enforcement or other parties except as required by valid legal process

Aggregate research use

Aggregate, anonymized query patterns may inform future research at the project’s discretion. Examples:

• “Users frequently ask about the relationship between dossier A and dossier C; the corpus should make this cross-reference more discoverable”
• “The refusal pattern for substance-dosing queries fires 50 times per week; this pattern is real and the dossier handling of substance use should perhaps include a more prominent direct-to-clinician redirect”
• “Queries about specific named living individuals are filtered out at high rate; this confirms the refusal pattern is necessary”

Aggregate use does not involve any individual user’s identifiable data. Where research using aggregate patterns is later published, the publication credits the corpus and discloses the aggregate-pattern methodology.

Account-bound use (if implemented)

If the AI reference tool implements an optional sign-in feature for users who want their query history available to them across sessions:

• Sign-in is optional; the tool works without sign-in
• Account creation requires only an email address (not a real name) and a password
• Account-bound query history is retained for the account’s lifetime plus 90 days after account deletion (the same window as the anonymous-query retention)
• Account deletion is available at any time via account settings; deletion removes the account email, password hash, and associated query history from operational storage
• Aggregate patterns derived before account deletion remain in aggregate form; the individual identifying data is removed

Children

The project does not knowingly collect personal data from children under 16. The published material is academic-rigor research about often-graphic religious, mystical, and esoteric subjects; it is not designed for or directed at children. If you are a parent or guardian and believe your child has provided personal data to the site, contact the publisher and the data will be removed.

Legal process and disclosure

The project operates under U.S. jurisdiction (4LULZ Press / Locked LLC, Montana). Standard legal process may require the project to retain or produce data:

• Valid subpoenas, court orders, or other lawful process will be reviewed for scope and legitimacy; the project will comply with valid process and will challenge process that is overbroad, lacking jurisdiction, or otherwise legally defective
• The project does not affirmatively cooperate with extralegal data requests
• The project does not respond to data requests from foreign governments except as required by treaty or U.S. legal process
• The project will notify affected users of legal process requesting their data when not legally prohibited from doing so

The project does not engage in voluntary disclosure to government agencies, advertisers, marketing partners, or third parties beyond the operational service providers named on this page.

Service providers

The project uses the following operational service providers, each with their own privacy policy:

• GitLab Pages — static-site hosting; standard server logs
• Cloudflare — DNS, CDN, web analytics (privacy-preserving analytics product)
• Kit (ConvertKit) — email subscription service
• Anthropic API — if the AI reference tool is deployed, query processing uses Anthropic’s Claude API; Anthropic’s API privacy and data-use policies apply to the inference call; the project’s instructions to Anthropic API include no-training-on-our-inputs flags where Anthropic offers them
• A vector database / RAG infrastructure provider — to be selected when the AI tool is built; will be disclosed here at that time

Each provider holds the data described in their respective policies. The project does not transfer subscriber email lists, query data, or analytics data between providers beyond what is necessary for the specific service.

Your rights

Depending on your jurisdiction, you may have specific rights regarding personal data the project holds about you. These may include:

• The right to know what data the project holds about you
• The right to access that data
• The right to correction of inaccurate data
• The right to deletion of data (subject to operational and legal retention requirements)
• The right to portability of data in a machine-readable format
• The right to object to certain processing

Exercise these rights by contacting the publisher through the standard contact channel at thefire.lol. The project will respond within a reasonable time and will not retaliate against rights-exercise requests.

Changes to this policy

When this policy changes, the change is reflected on this page with an updated last-modified date. Material changes (changes that expand data collection, lengthen retention, or add new categories of use) are announced via the newsletter so existing subscribers are notified. Non-material changes (clarifications, typo corrections, link updates) are made without separate announcement.

Contact

Questions about this policy, requests to exercise rights described above, or specific concerns about the project’s handling of your data: through the publisher’s standard contact channel at thefire.lol.

Last updated: 2026-05-14. Maintained by 4LULZ Press / thefire.lol.