Privacy Policy

Chess and Stilz Co. LLC

Effective Date: October 12, 2026 | Last Revised: October 31, 2026

Table of Contents

1. About This Policy and Who It Covers
2. Our Product in Brief
3. Information We Collect
4. How We Use Information
5. The Reader Dial — Behavioral Inference Disclosure
6. The Reader Recap — Parent Communication
7. AI-Generated Story Content
8. How We Share Information — Parties
9. Data Retention
10. Security
11. Parent Rights Under COPPA
12. California Residents — CCPA/CPPA
13. Changes to This Policy
14. Contact Us

1. About This Policy and Who It Covers

Chess and Stilz Co. LLC ("Chess and Stilz," "we," "us," or "our") operates an interactive storytelling subscription service for children age 9 and older (the "Service"), accessible via browser on any modern device at https://chessandstilz.com. No download or installation is required. This Privacy Policy describes how we collect, use, share, and protect personal information in connection with the Service.

This Service is directed to children. We knowingly collect personal information from and about children under 13. The Children's Online Privacy Protection Act (COPPA), 15 U.S.C. § 6501 et seq., and its implementing rule, 16 CFR Part 312 (including the April 2026 amendments), apply and are accepted. We do not allow any child to create an account, provide an email address, or make any purchase. The account holder is always a parent or legal guardian ("Parent"), and all consent and control flows through that Parent.

This Policy also addresses the rights of California residents under the California Consumer Privacy Act (Cal. Civ. Code § 1798.100 et seq.) and the California Privacy Rights Act, collectively referred to here as CCPA/CPPA.

If you have questions about this Policy or our data practices, see Section 14.

2. Our Product in Brief

The Service delivers personalized interactive stories using an "authored skeleton, generative delivery" architecture: Chess and Stilz authors story frameworks and a hard-coded content safety floor; AI model providers generate prose at runtime within those frameworks. The child navigates by tapping pre-authored choices — there is no free-text input anywhere in the product, no chat function, and no communication path between the child and any real person.

The paying customer is always the Parent. The child has no account, login credentials, email address, or purchasing capability. The parent's account has two zones: the Kid Zone — the child-facing story reading experience — and the Parent Zone — the parent-only settings and management area, protected by the Parent PIN (a separate credential from the account password).

Minimum age: 9.

3. Information We Collect

3.1 Information Collected From the Parent at Signup

3.2 Information Collected During Child Profile Setup (Parent-Entered)

The following items are entered by the Parent, not the child. The child never sees or interacts with a free-text field anywhere in the Service.

Character name. A single token entered by the Parent during child profile setup. Spaces are blocked at input — the field cannot accept a first-and-last name together. Hyphens and apostrophes are permitted (e.g., Star-Captain, D'Artagnan). The UI prompt reads: "What is your character called in our stories? Nicknames and fantasy names encouraged." The character name is used to personalize story prose and is included in the story-state payload sent to AI model providers (see Section 7). A Parent PIN is required to edit it.

Reading age. An integer, Parent-set, default 9. Drives the vocabulary register of generated prose.

Bravery tier. Level 1, 2, or 3, Parent-set. Drives peril intensity within stories.

Dark Content toggle state. A single toggle gates the Dark Content category — when ON, it covers all worlds in that category, current and future. A Dark Content category is a Chess and Stilz editorial classification carrying a defined safety floor and editorial register; specific story worlds are placed into a category by Chess and Stilz's editorial judgment. The Service includes one parent-opt-in Dark Content category. At launch, no worlds are placed in this category; worlds with darker editorial register may be added to the category in future releases. Default OFF for every child regardless of age or editorial tier. Delivery of content from any Dark Content world is subject to a two-lock gate: (1) the child's editorial tier must match the world's minimum tier, AND (2) the Parent must have toggled Dark Content ON. Both conditions must be simultaneously true — the editorial tier advancing alone is not sufficient.

Chess and Stilz may add additional worlds to an existing Dark Content category over time, provided the new world matches the category's existing safety floor and editorial register. Parents consenting to a category consent to all worlds within it; additions of new worlds within an already-consented category do not constitute a material change requiring new consent. Material changes to a category's safety floor or editorial register would be treated as a material change under Section 13.

3.2.1 World-Entry Preference Choices (Child-Selected, Set Once Per World)

At the first entry into each story world, the child selects from pre-authored answer options (no free-text input) to indicate two persistent preferences for that world:

• A reaction-style preference (e.g., how the child prefers stories to unfold within that world)
• A character-affinity preference (e.g., which character archetype the child wants to follow most closely)

These are stored declarations, not behavioral inferences. The child directly selects from pre-authored options; the system does not analyze behavior to derive them. They are stored with the child's profile and included in the story-state payload sent to AI model providers on every generation call for the relevant world. They persist across sessions until the child profile is reset (see Section 11.1) or deleted. They cannot be edited by the Parent.

3.3 Behavioral Telemetry Collected During Child Reading Sessions

The Service automatically records the following data during each reading session:

• Story choice picks — which fork was taken at chapter-ending decision points and which inline choices were made
• Choice latency — time elapsed between choice presentation and the child's selection
• Scene dwell time — time spent on each scene or passage
• Character engagement signals — per-character depth-of-engagement indicators
• Pace signal — kinetic (fast-moving) vs. atmospheric (lingering) reading pace
• Thread decline state — whether a given active story thread was declined
• Session start and end timestamps

This telemetry is used to operate the Reader Dial (see Section 5) and to generate the Reader Recap (see Section 6).

3.4 What We Do Not Collect

We do not collect:

• The child's actual birth date
• The child's real first or last name
• Any free-text input from the child
• Voice, audio, image, or video data
• Biometric information
• Location data
• Device-level identifiers beyond what is required for session continuity
• Cross-app or cross-site behavioral data
• Data from third-party analytics services with access to child telemetry
• Reading-level inferences beyond the parent-set reading age

4. How We Use Information

We do not sell personal information. We do not use personal information for advertising. News about the Service is separate from the Reader Recap. It can be turned off at any time using the unsubscribe link in any such email, and turning it off does not affect the Reader Recap, your account notices, or billing notices. We do not share child information with third parties for any purpose other than operating the Service, as described in Section 8.

We do not condition your child's participation in the Service — or in any activity within the Service — on disclosure of more personal information than is reasonably necessary for that activity.

5. The Reader Dial — Behavioral Inference Disclosure

The Reader Dial is the system through which Chess and Stilz personalizes the story experience within the parameters you have set. It operates in three layers:

Layer 1 — Parent-set guardrails. The Reader Dial begins with the inputs the Parent sets directly: reading age, bravery tier, safety settings, and the Parent's Dark Content toggle state. These values are sent to AI model providers with every story generation call so that prose stays inside the Parent's guardrails. Past-ratings — whether a story segment was rated favorably or unfavorably in prior reading sessions — are the simplest form of pattern adaptation the system uses, and are included in Layer 1.

Layer 2 — Child-Contributed Signals. During an active reading session, the Service derives a single signal — the reading lean — that reflects whether the child's engagement is leaning kinetic (faster-paced, action-forward) or atmospheric (slower, character- and setting-forward). This signal is included in the story-state payload sent to AI model providers at runtime, alongside Layer 1 values. The child's world-entry preference choices, which are stored declarations rather than in-session inferences, are described in Section 3.2.1.

Layer 3 — Deeper behavioral inference. Additional behavioral inference based on aggregated session telemetry (dwell patterns, choice latency, character engagement signals) is not active in the current version of the Service. If Layer 3 is introduced in a future release, fresh verifiable parental consent (VPC) will be obtained before activation for any child profile, consistent with COPPA's advance notice requirements for behavioral profiling.

How it is used. The Reader Dial is used exclusively to tune story register, pacing, character emphasis, and story selection — always within the hard-coded safety floor and always within the Parent-set guardrails of reading age, bravery tier, and Dark Content toggle state. The Reader Dial never overrides a Parent setting.

How it is not used. The Reader Dial is never used for advertising, never sold, and never shared with third parties for their own purposes. It is never shown to the child. It is never expressed as a numerical score or rating to anyone.

Dial-derived hints and AI model providers. At runtime, two signals derived from the Reader Dial — a pace hint (kinetic vs. atmospheric reading pattern) and character-depth hints (which characters the child engages with most deeply) — are included in the story-state payload sent to AI model providers to generate the next story segment. These are derived signals; the underlying per-event telemetry logs (choice latency, dwell timestamps, session records) are never transmitted to AI providers. AI model providers receive these hints as inputs to a single generation call. The provider retains this input for up to 30 days under its standard commercial data-handling terms — used only to operate and secure the service, never to train the provider's models — and then permanently deletes it, subject to the rare flagged-content exception described in Section 7.3(5). This transmission is an operational input to story delivery, not a disclosure or sharing in the commercial or analytical sense.

Parent visibility. Parents receive a plain-English narrative summary of the Reader Dial in their Reader Recap (see Section 6). The underlying numerical inference is not surfaced.

COPPA disclosure. The Reader Dial constitutes behavioral profiling of a child under 16 CFR Part 312, as amended April 2026. This disclosure, and a corresponding disclosure in the Direct Notice to Parents, satisfy the COPPA requirement for advance notice of behavioral profiling. Verifiable parental consent (VPC) is obtained before the Reader Dial is activated for any child profile.

Retention. Consistent with 16 CFR § 312.10, Reader Dial inferences are not retained beyond what is reasonably necessary for the story-tuning and Reader Recap purposes described above. Reader Dial inferred state is deleted immediately and in full upon deletion of the associated child profile.

Aggregate editorial use. Anonymized aggregate signals may inform Chess and Stilz's editorial authoring decisions over time, and may be shared in aggregate form with educational research and curriculum development partners. Any aggregate data shared with such partners is subject to written agreements prohibiting re-identification, commercial use for advertising, and further sharing. These signals are drawn only from cohorts large enough that no individual reader's behavior is identifiable, are not linked to any individual child's profile, and are never used to communicate with anyone other than that child's Parent through the Reader Recap.

6. The Reader Recap — Parent Communication

The Reader Recap is a communication delivered each time a chapter is finished to the Parent's account email address and accessible in the parent dashboard (the "Parent Zone") within the Service. It has two parts:

Part 1 — The Letter. A narrative written in the voice of the in-world guide. The Letter describes your child's engagement and story progress in plain English — no metrics, no numerical scores, and no grading language.

Part 2 — The Snapshot. Factual engagement data (session count, total time spent during the period) alongside one-click adjustment prompts for bravery tier and locked content categories.

The Reader Recap is the only channel through which behavioral data about your child is communicated to you as a Parent — it flows only to your account email address and Parent Zone. (Dial-derived hints transmitted to AI model providers as operational inputs to story generation are described in Section 7 and are distinct from outward communication.) This is a Parent-facing information flow encouraged by COPPA. It is disclosed in the VPC consent screen and in this Policy.

Parent controls. Parents may disable the Reader Recap entirely through the Parent Zone. Every edition remains available in the Parent Zone. Disabling the Recap does not affect service delivery or the Reader Dial's operation. Parents may clear Recap history along with accumulated reading data without deleting the account or revoking consent, using the Reset Child Profile action in Parent Zone → Privacy & Data (see Section 11).

7. AI-Generated Story Content

7.1 How AI Is Used

Chess and Stilz uses AI to generate some story prose at runtime. Where it is used, the prose is produced by an AI model operating within an authored story framework (the skeleton of characters, world rules, and plot structure created by Chess and Stilz authors) and within a hard-coded content safety floor that applies at all ages. AI generation is disclosed at signup, in this Privacy Policy, and within the Service.

7.2 AI Model Providers

AI model providers are US-headquartered, tier-one providers. Chess and Stilz does not name specific providers in this Policy — the Service architecture is provider-agnostic, and specific models are evaluated and updated over time. Chess and Stilz commits to the category of provider and the posture commitments below, so that this Policy does not require amendment each time a provider or model changes. An internal audit record identifies the specific provider(s) in use at any given time.

7.3 Posture Commitments (Stable Across Model Changes)

These commitments apply regardless of which specific AI model is in use:

1. No child data used for model training. No information about your child is used to train, fine-tune, or otherwise improve any AI model provider's models.
2. Defined payload to providers. The story-state payload transmitted to AI model providers consists of: (a) the character name, (b) the current story state, (c) the choice the child most recently made, (d) Reader Dial values — the Parent's guardrail settings (reading age, bravery tier, safety settings, and Dark Content toggle states) and past-ratings, (e) the child's world-entry preference choices (see Section 3.2.1), and (f) the reading lean signal — the in-session kinetic vs. atmospheric engagement lean. No parent identity information, no raw behavioral telemetry, no deeper behavioral inference (Layer 3 is not active in the current version), and no demographics are sent to any AI model provider. These payload items are operational inputs to story generation and are subject to the retention and non-training commitments described in commitment 5.
3. Disclosed generation. AI-generated content is disclosed at every touchpoint described above.
4. Filtered output. All AI-generated prose and imagery is passed through a two-pass content safety filter (prompt-side and output-side) before being delivered to the child's screen. The filtering system is also AI-based. The two-pass filter is Chess and Stilz's safety model. Bravery tier, Reader Dial state, Dark Content toggles, and world-entry preference choices are taste and register controls — they shape vocabulary, peril intensity, character emphasis, and story register within the safety floor. They are not safety controls and have no effect on what categories the filter blocks. No combination of user settings can permit content that the safety floor filters. AI-generated prose is produced fresh for each child reader at runtime within Chess and Stilz's authored frameworks, and may resemble prose generated for other readers in style or structure given shared frameworks and shared model providers; Chess and Stilz does not warrant that any segment of generated prose is unique to a particular reader.
5. Standard-terms retention; no training use; flagged-content exception. Our AI model provider retains the story-state payload for up to 30 days under its standard commercial data-handling terms and then permanently deletes it. This retention period reflects the provider's standard terms of service — it is not a bespoke zero-data-retention arrangement negotiated for Chess and Stilz, and Chess and Stilz does not represent that any AI model provider retains no data. During this 30-day window, the provider does not use any data we transmit to train, fine-tune, or otherwise improve its models, and does not use it for any purpose other than generating the requested story response and operating and securing its service.

Flagged-content exception. In the rare event that a specific story-generation request is flagged by the provider's own automated safety-review systems as an apparent violation of the provider's usage policies, the provider may retain that flagged input and output for up to 2 years, and an associated safety classification score for up to 7 years, for the provider's own safety-enforcement purposes. This exception can apply even where the content fully complies with the two-pass content safety architecture described in commitment 4 above and Section 6 of our Terms of Service — it reflects the provider's own independent, automated classification system, not a judgment about your child's specific story or a failure of Chess and Stilz's own filtering. Chess and Stilz has no visibility into individual flagging determinations at the provider and no ability to expedite deletion of a flagged record.

Chess and Stilz maintains an internal record of the providers in use and their applicable data-handling terms at any given time, and will update this Policy if the retention period, training posture, or flagged-content exception of any provider materially changes.

7.4 Story Imagery

Story imagery in the Service consists of character portraits and illustrations from a curated library authored by Chess and Stilz using AI-assisted tools and editorially reviewed before publication. No child personal information is transmitted to any image generation service. AI-assisted authoring is disclosed in our AI Tools / Credits page at https://chessandstilz.com/credits.

8. How We Share Information — Parties

Chess and Stilz shares personal information only with the following named parties, solely for the purposes of operating the Service. We do not share personal information with advertising networks, data brokers, analytics partners with access to child telemetry, social media platforms, or content syndication partners.

A current list of authoring-time AI tools (used in producing story frameworks, character art, and editorial materials before publication) is published at https://chessandstilz.com/credits and is updated when tools change.

No other third-party sharing. Chess and Stilz does not sell personal information. Chess and Stilz does not share personal information with any party not listed above, except as required by law (e.g., in response to a valid legal process, or to protect the safety of a child).

9. Data Retention

Three data lifecycle paths:

Path 1 — Explicit deletion (COPPA right to delete). A Parent who uses Parent Zone → Privacy & Data to delete a child profile or the full account triggers an immediate, permanent hard-delete of all associated data in Chess and Stilz's own systems — telemetry, Reader Dial state, Recap history, and the account record. There is no freeze period, no undo window, and no email-based alternative. This action is available at any time, including concurrent with subscription cancellation.

Path 2 — Subscription cancellation. A Parent who cancels the subscription without using the deletion options in Parent Zone → Privacy & Data initiates a 90-day save-back hold. During the hold, Chess and Stilz stops all processing: no new telemetry is collected, no Reader Dial updates occur, and no Reader Recaps are generated. Reactivation emails with a one-click return link are sent at day 7, day 30, and day 75. If the account is reactivated before day 90, service resumes from the saved state. If no reactivation occurs, all data is permanently deleted from Chess and Stilz's own systems at day 90. Cancellation alone is not a deletion request — a Parent who wants data deleted immediately uses Path 1.

Path 3 — Editorial Reset (not a deletion right). A Parent who uses the Reset Child Profile action in the Parent Zone (see Section 11.1) triggers immediate, permanent deletion of session telemetry, Reader Dial inferred state, persistent world-entry preference choices, and Reader Recap history in Chess and Stilz's own systems. The child profile shell — character name, reading age, bravery tier, and Dark Content toggle states — is preserved. Reset is an editorial product feature, not a COPPA deletion right.

Scope of these paths — AI provider retention is separate. Paths 1 through 3 above describe deletion from Chess and Stilz's own systems. Story-generation requests already sent to our AI model provider are retained independently, under the provider's own terms described in Section 7.3(5): every request — regardless of whether it was ever flagged — is held by the provider for up to 30 days under its standard retention window, and a request flagged by the provider's own automated safety-review systems may be retained for up to 2 years (with the associated safety classification score retained for up to 7 years). In either case, that provider-held copy is not necessarily deleted by any action taken under Paths 1–3 — Chess and Stilz has no visibility into individual retention or flagging determinations at the provider and no ability to expedite deletion there.

10. Security

Chess and Stilz uses industry-standard technical and organizational measures to protect personal information, including:

• Encryption in transit for all data moving between the Service, our servers, and subprocessors
• Encryption at rest for all data stored in our managed database infrastructure (Fly.io)
• No stored card data — Chess and Stilz never receives or stores payment card numbers; Stripe holds all payment credentials
• Parent PIN gating for all sensitive child-profile actions in the Parent Zone (editing character name, downloading data, deleting profiles, deleting accounts)

No security system is impenetrable. In the event of a security incident affecting personal information, we will notify affected users as required by applicable law.

11. Parent Rights Under COPPA

Federal law gives you — as a Parent — the right to:

• Review personal information we have collected about your child
• Correct inaccurate information
• Refuse further collection or use of your child's information
• Request deletion of your child's information

Chess and Stilz does not condition your child's participation in the Service — or in any activity within the Service — on disclosure of more personal information than is reasonably necessary for that activity.

11.1 Self-Service Actions in the Parent Zone

All three core COPPA rights are available as self-service actions in Parent Zone → Privacy & Data. Each action is PIN-gated, automated, and completed immediately on confirmation. No email channel or manual SLA is required.

Action 1 — Download my child's data. Exports a JSON file downloaded from the Parent Zone containing: the account record, child profile settings, all session telemetry, Reader Dial state, and Reader Recap history. This satisfies the right to review.

Action 2 — Delete child profile. Permanently removes: the child profile, all session telemetry, Reader Dial state, and Reader Recap history. The Parent's account and subscription remain active. Deletion constitutes revocation of COPPA consent. Chess and Stilz immediately and permanently ceases processing the child's personal information.

Action 3 — Delete account entirely. Removes the account and all associated data. The Stripe subscription is cancelled in accordance with subscription terms. Access ends immediately. No automatic refund is issued.

Note on subscription cancellation. Actions 2 and 3 above trigger immediate, permanent deletion — this is the COPPA right-to-delete path. Cancelling your subscription through billing settings without using one of these actions is different: it initiates a 90-day data hold before permanent deletion. If you want your child's data deleted immediately, use Action 2 or Action 3. See Section 9 for the full lifecycle.

11.1.4 Editorial Reset (Not a COPPA Right)

The Parent Zone also offers a Reset Child Profile action, which clears accumulated reading data — session telemetry, Reader Dial inferred state, persistent world-entry preference choices, and Reader Recap history — while preserving the child profile shell (character name, reading age, bravery tier, and Dark Content toggle states). Reset does not revoke COPPA consent, does not delete the child profile, and does not affect subscription status. It is an editorial product feature that lets the reading experience start fresh. Parents seeking deletion of personal information should use Action 2 or Action 3 above. Reset is PIN-gated and requires explicit confirmation that progress and Reader Dial state cannot be recovered.

11.2 Right to Correct

Parents may edit the child's character name (PIN-gated) and all Parent-set fields — reading age, bravery tier, action and conflict, scary or spooky elements, per-setting locks, and Dark Content toggles — directly in the Parent Zone at any time. No support request is required.

11.3 Contacting Us

To exercise any right by email, or if you have questions about this Policy or our data practices, contact us at hello@chessandstilz.com (see also Section 14).

12. California Residents — CCPA/CPPA

This section applies to California residents and supplements the information elsewhere in this Policy.

Categories of personal information collected. We collect the categories described in Section 3: identifiers (parent name, email), commercial information (payment records via Stripe), and internet or other network activity (session telemetry, Reader Dial behavioral inferences, story choices, and world-entry preference choices). We do not collect sensitive personal information beyond what is described in this Policy.

Right to know. You have the right to know the categories of personal information we collect and how it is used. This Policy provides that information.

Right to delete. You have the right to request deletion of personal information we hold about you or your child. See Section 11.1 for self-service deletion options.

Right to correct. You have the right to request correction of inaccurate personal information. See Section 11.2.

Right to opt out of sale or sharing. Chess and Stilz does not sell personal information and does not share personal information for cross-context behavioral advertising. No opt-out mechanism is required; none is needed.

Right to limit use of sensitive personal information. We do not use sensitive personal information for purposes beyond those permitted by the CCPA/CPPA without first presenting a right to limit.

Right to non-discrimination. Exercising any privacy right described in this Policy will not result in denial of service, a different price, or a different level of service quality.

Authorized agent requests. An authorized agent may submit a rights request on your behalf by contacting us at hello@chessandstilz.com. We may verify the agent's authority before processing the request.

Response timeline. We will respond to verified requests within 45 days. If we require additional time (up to 90 days total), we will notify you of the extension within the initial 45-day period.

13. Changes to This Policy

We may update this Privacy Policy from time to time. Material changes — meaning changes that affect the nature of data collected, the purposes for which it is used, or the rights available to you — will be communicated by email to your account address before the change takes effect. The notice will state the effective date of the updated Policy.

Your continued use of the Service after the notice period constitutes your acceptance of the updated Policy. If you do not accept a material change, you may delete your account before the change takes effect.

Non-material clarifications (grammar, formatting, reorganization without substantive change) may be made without advance notice. The "Last Revised" date at the top of this Policy reflects the date of the most recent update.

14. Contact Us

Questions about this Privacy Policy, requests you prefer to submit by email rather than through the Parent Zone, or reports of suspected privacy incidents:

Chess and Stilz Co. LLC 176 S. Sunset Dr., Vineyard, UT 84059 +1 801-477-7675 hello@chessandstilz.com

We aim to respond to all privacy inquiries within five business days.

Chess and Stilz Co. LLC | hello@chessandstilz.com | Last revised: October 31, 2026