Privacy Policy

Last updated: October 23, 2026

Legal entity: Orange Zest Inc., Ontario, Canada

Website: https://trove.is

Full Privacy Policy

1. Who We Are

Trove ("we," "our," or "us") refers to Orange Zest Inc., based in Ontario, Canada. We operate the Trove platform, accessible at https://trove.is and its associated applications.

This Privacy Policy explains how we collect, use, store, and protect your information when you use Trove. By using our service, you agree to the practices described here.

2. Information We Collect

Account Information

When you create an account, we collect:

• Your display name
• Email address
• Pronouns (if provided)
• Birthday, birthplace, places you have lived, and other profile details you choose to provide
• Authentication data through our identity provider (Google or Apple sign-in, or email/password)

Behavioral and Interaction Data

When you play through Trove scenarios ("Tangles"), we collect:

• Every choice you make within a scenario
• Your written responses in interactive conversations
• The sequence and pattern of your decisions
• We may also collect interaction signals such as response timing and revision patterns

This behavioral data is the core of how Trove works — it's what allows us to generate personalized insights about your patterns.

Optional Public Sources in Trove Profile

For Bluesky and public website Profile sources, Trove shows you a preview before you authorize a one-time import or daily refresh. These sources include original public text posts from a Bluesky profile or text from one public website page you provide. Social Read can also retrieve public Instagram or LinkedIn profile text and recent original post text through Apify and its actor services. For those requests, Apify receives the public profile URL; Trove receives selected bio and post text, links, and dates. We do not ask for your social account password. We store the imported text, source URL, and available publication and retrieval dates with your private Profile. A public URL does not verify that you own or wrote its content.

We use the imported text itself to provide your Profile read. For Trove model improvement, research, or partner sharing, we may separately create patterns and insights derived from it. Before those other uses, we remove or exclude identifying details and passages traceable to the original source. Research outputs are anonymized and aggregated; partner data is de-identified and aggregated, as described below. We do not use the original imported text for these other purposes. If that changes, we will update this policy and ask for consent where required.

Voice Input

Voice input is optional. The installed app uses the device's speech recognition service, which can process audio through Apple or the operating system's speech provider. Trove does not require on-device-only recognition. The app does not persist a voice recording; you can edit the transcript before submitting it. Web experiences with realtime transcription send audio to OpenAI. When you submit a transcript, its text is processed and stored like a typed response and can be included in the AI feature you use.

Automatically Collected Information

When you use Trove, we automatically collect:

• Device and browser information
• IP address (used for approximate location and analytics)
• Pages visited and actions taken within the app
• Session data and interaction events
• Error and performance data

3. How We Use Your Information

We use the information we collect to:

• Generate and adapt interactive scenarios in real time based on your choices
• Produce personalized behavioral insights and results
• Use public source material you choose to import to support your private Trove Profile read, and check daily for updates only if you enable that option
• Maintain your account and allow you to access your results over time
• Understand how people use Trove and improve the experience
• Study patterns in how people engage with Trove — both for our own product work and, sometimes, for research outputs like papers, talks, or academic collaborations. This always uses anonymized, aggregated data — see Section 7 below for the full breakdown
• Personalize your future experiences using patterns from across the scenarios you've played, and develop and improve Trove's own models and personalization systems — using de-identified data wherever practical (see Section 4)
• Communicate with you about the service (e.g., updates, support)
• Detect and prevent misuse of the platform

4. How AI Powers Trove

Trove uses artificial intelligence to create interactive experiences and generate personalized insights. Here's how it works:

What AI does in Trove:

Trove uses large language models to generate scenario content, adapt stories based on your choices, and develop personalized insights. Newer Trove Profile features use DeepSeek models hosted by DeepInfra through OpenRouter. Anthropic and OpenAI remain in use for other experiences and services, including audio transcription where enabled.

What data is sent to AI providers:

When you play a scenario, the following data is sent to our AI providers for processing:

• Your display name and pronouns
• Your choices and responses within the current scenario
• The conversation history within the current scenario

When we generate or revise your private Trove Profile read, we send the selected AI services the relevant Profile evidence and feedback. Depending on the feature, this can include your name, birthday and place history, original questionnaire answers, conversation responses, prior insights, and imported source text, links, and dates. OpenRouter routes current DeepSeek requests to DeepInfra; selecting a DeepSeek model does not mean we send these requests directly to DeepSeek's own API.

What we do not send as separate fields to AI providers:

• Your account email address (text or feedback you provide may still contain personal details)
• Your authentication credentials

Personalization across scenarios:

As Trove develops, we may use patterns from across the scenarios you've played to personalize and adapt future experiences for you. When we do, relevant parts of your behavioral profile may be included in the data sent to our AI providers to generate your personalized content. This profile is built from your own gameplay and is used to tailor your experience.

AI provider data practices:

• Anthropic: Processes inputs and outputs for the experiences that use its API. Retention depends on the service, contractual settings, and safety or legal exceptions. See Anthropic's retention documentation.
• OpenAI: Processes inputs, outputs, and audio for the services that use its API. Retention depends on the endpoint, account settings, and safety or legal exceptions. See OpenAI's data controls documentation.
• OpenRouter: Routes requests to the selected model host and records operational metadata. Input/output logging and data-use options depend on account settings. See OpenRouter's data collection documentation.
• DeepInfra: Hosts the DeepSeek models used by current Profile features. Its data handling is governed by the applicable service terms and privacy policy. See DeepInfra's terms and privacy policy.

Retention and data-use terms differ between providers and processing routes. We do not promise a uniform deletion period or zero retention across these services. Trove's own uses of data are described below.

Improving Trove's own models and personalization:

We may use your behavioral data — your choices, responses, and patterns, including patterns and insights derived from optional Profile sources — to develop, train, and improve Trove's own models and personalization systems: the systems that generate your scenarios and insights and tailor them to you. For this internal work we use your data in de-identified form wherever practical. If we ever wanted to train a model on data that could be linked back to you, or build a model whose outputs draw on your data to serve other users, we'd ask for your explicit consent first.

5. Third-Party Services We Use

We use the following third-party services to operate Trove:

We do not share data that identifies you with advertisers, data brokers, or marketing partners. We may share or sell de-identified, aggregated data — processed so it can't be linked back to you — with research or commercial partners. This de-identified data does not include your name, email, account, or anything that could reasonably be used to re-identify you.

That may include aggregated patterns derived from optional Profile sources, but not the imported posts or page text itself.

6. Analytics and Tracking

We use PostHog for product analytics to understand how people use Trove, including page views, scenario completions, and feature usage. Earlier web experiences may associate analytics with your account email. The native Profile app uses your sign-in account identifier, platform and build information, and fixed usage or error codes. Its telemetry does not include your answer text, notes, transcripts, imported source text, email, or generated readings; session replay is disabled in that app.

We use cookies and similar technologies to maintain your login session and enable core functionality. The public landing website also includes Google Analytics and the Meta Pixel for website measurement, including page views, download-button clicks, and waitlist submissions. These services can receive page and event information, browser identifiers, and network metadata. These website integrations are separate from the native app telemetry described above.

7. Research Use

From time to time, we publish or present what we learn from how people use Trove — for example, in research papers, conference talks, blog posts, or collaborations with academic researchers. We think the patterns Trove surfaces are interesting beyond just our own product, and sharing what we learn helps both Trove and the broader field of behavioral research.

When this applies: the research-use approach in this section covers data collected from February 2026 onwards (when this privacy policy was first established).

Trove is a Canadian company, so we follow Canada's privacy rules here. The short version: we only use anonymized data for research. "Anonymized" means the strong sense — processed into a form that can't be linked back to you, even by us.

If we ever wanted to use data that could be linked back to you for a research project, we'd ask you first — for that specific study, not buried in this policy.

Here's exactly how the anonymized-research process works:

What we use:

• Aggregated patterns from gameplay — for example, the distribution of how groups of people responded to a particular scenario
• Anonymized, aggregated patterns derived from optional Profile sources or insights, without the imported posts, page text, or traceable excerpts
• Anonymized excerpts of free-text or chat responses, where they help illustrate a finding
• Anonymized scoring outputs

What we strip out before any of this leaves our systems:

• Your name, display name, and pronouns
• Your email address and account ID
• Any internal database IDs we use to tie records to your account
• Any direct identifiers you wrote into a free-text response (for example, a real person's name) — we either redact these or exclude the response entirely
• Combinations of fields that could plausibly re-identify someone (for example, very rare scenario-completion patterns from a small subgroup)

What we never share externally:

• Your account record or raw account data
• Full transcripts of your gameplay tied to you
• Anything that could reasonably be used to re-identify you

Quoted free-text responses:

If we ever quote something you wrote, we apply two checks first: (1) the response contains no direct identifiers (names, emails, phone numbers, employers, schools, addresses), and (2) the response isn't so specific that it could re-identify you when combined with publicly known facts. If a response fails either check, we either redact the identifying parts or leave it out entirely.

Academic collaborations:

When we work with academic researchers, they have to follow their own institution's research-ethics rules — that's a real review process with a board that signs off on the study. If a specific study ever asks for more from you than what's described here (for example, participation in a separate study, or use of data that could be linked back to you), we'll come to you directly first.

Live in Quebec?

Quebec has its own privacy law (Law 25) that gives you extra rights around how your information is used for purposes beyond running the service. The opt-out below works for you the same way — same email, same process. If you'd like us to confirm exclusion in writing, just ask in your email and we will.

Opting out:

Anything we've already published stays out there — but it only ever used aggregated, anonymized data. Going forward, you can opt out of research use of your data at any time by emailing [email protected] with the subject line "research opt-out." Opting out won't affect your account, your results, or your ability to use Trove. Once you opt out, we exclude your data from any future research output. This research opt-out covers research outputs specifically — papers, talks, and academic collaborations. It is separate from the de-identified, aggregated data we may share with partners (Section 5), which never contains anything that could be linked back to you.

8. How We Store and Protect Your Data

Your data is stored using Convex, a cloud database service. We use industry-standard security practices including encrypted connections (TLS) and access controls to protect your information.

Your behavioral data (scenario choices and results) is stored in our database for as long as you maintain your account, so you can access your results over time.

9. Data Retention

• Account data: Retained for as long as your account is active.
• Behavioral data (choices and results): Retained for as long as your account is active.
• Optional Profile sources: Imported text can remain with your account until you remove its source or delete your account; a refresh can also replace or clear it. Pausing daily refresh stops new collection but keeps existing material. Removing a source deletes its imported text and removes source evidence and dependent insights from saved reads. Feedback you wrote may remain. Derived patterns already included in separate research or model-improvement datasets may remain as described below. Backup copies may persist for a limited period.
• AI provider data: Retention depends on the service and processing route. See the provider-specific information in section 4; a 30-day limit is not shared by every service.
• Analytics data: Retained by PostHog in accordance with their data retention policies.
• Research datasets: Anonymized, aggregated datasets used for research are kept separately from your account and may be retained for the life of the research project, even if you later delete your account. These datasets do not contain identifying information. If you opt out of research use, your data is excluded from any new research datasets going forward.
• Model and personalization datasets: De-identified datasets used to train and improve Trove's own models are kept separately from your account and may be retained even if you later delete your account. These datasets do not contain identifying information.
• Deleted accounts: When you request account deletion, we will remove your personal information and behavioral data from our active systems. Some data may persist in backups for a limited period.
• Deletion protection: After account cleanup, we retain a one-way hash of the sign-in identifier to block delayed requests from recreating deleted records. It contains no profile, answers, name, or email. This security marker remains until an operator can independently verify that removing it cannot restore access or recreate the account.

10. Your Rights and Choices

Access and Deletion:

In the installed app, open Account → Settings & privacy and choose Download my data or Delete account. These actions can require recent sign-in verification. You can also request access or deletion help by emailing [email protected]. When you request deletion, we will remove your account, profile information, scenario choices, and behavioral results from our active systems.

In Trove Profile, you can inspect a source before import, choose a one-time import without automatic collection or enable daily refresh, pause or resume refresh, and remove the source and its imported text. Imported material may inform later reads until you remove it.

Opt Out of Analytics:

On the website, you can use browser privacy settings or extensions to block analytics services, including PostHog, Google Analytics, and the Meta Pixel. Browser settings do not control native app telemetry. Contact [email protected] about privacy choices in the app.

Opt Out of Research Use:

You can opt out of having your data used in research outputs (papers, talks, academic collaborations) by emailing [email protected] with the subject "research opt-out." This won't affect your account, your results, or your ability to use Trove. See Section 7 for details on what research use involves.

11. Age Requirements

Trove is intended for users aged 16 and older. We do not knowingly collect information from anyone under 16. If you are a parent or guardian and believe your child under 16 has provided us with personal information, please contact us and we will take steps to remove that information.

12. Data Transfers

Trove is operated by Orange Zest Inc. in Ontario, Canada. Our service providers (including database, AI, and analytics services) may process your data in the United States or other countries. By using Trove, you consent to the transfer and processing of your information in these jurisdictions.

13. California Residents (CCPA)

If you are a California resident, you have the right to:

• Know what personal information we collect, use, and disclose
• Request deletion of your personal information
• Not be discriminated against for exercising your privacy rights

We do not sell your personal information. We may share or sell de-identified or aggregated information, which is not considered personal information under the CCPA because it cannot reasonably be linked to you.

To exercise these rights, contact us at [email protected].

14. European Users (GDPR)

If you are in the European Economic Area, you have additional rights including the right to access, correct, delete, or port your data, and to object to or restrict certain processing. To exercise these rights, contact us at [email protected].

Our legal basis for processing your data is your consent (by using the service) and our legitimate interest in operating and improving Trove.

15. Canadian Users (PIPEDA and Quebec Law 25)

Trove is operated from Canada (Orange Zest Inc., Ontario) and complies with Canada's federal privacy law (PIPEDA) as well as provincial privacy laws including Quebec's Law 25 where applicable. You have the right to access your personal information, request corrections, withdraw consent for its use, and request deletion. To exercise these rights, contact us at [email protected].

Research use. Section 7 describes how we use your data for research. If we ever wanted to use information that could be linked back to you for research, we'd ask you first — for that specific study, not buried in this policy.

Quebec residents. Quebec's Law 25 gives you additional rights, including the right to know when a decision about you is made by an automated system, the right to get a copy of your data to take elsewhere, and the right to give or withhold consent separately for distinct purposes. Section 7 (Research Use) describes how we handle research-specific consent for Quebec residents.

16. Changes to This Policy

We may update this Privacy Policy from time to time. We will notify you of significant changes through the service or by email. Continued use of Trove after changes constitutes acceptance of the updated policy.

17. Contact Us

If you have questions about this Privacy Policy or your data, contact us at:

Email: [email protected]

Entity: Orange Zest Inc., Ontario, Canada