Wilds.ai Privacy Policy

Effective Date: April 11, 2026

What We Collect

Account information: Email address, display name, hashed password for email/password accounts, linked OAuth identities, and account settings.
Billing information: Stripe customer ID, subscription status, and purchase history. We do not store full payment card numbers.
Content you create: Stories, worlds, companion configurations, session messages, companion conversations, lore, assets, and other creative content you submit to the service.
Operational and analytics data: Session starts and ends, heartbeat playtime, favorites, shares, search and browse behavior, multiplayer joins, compiler runs, model/provider usage, token counts, latency, and other product-usage telemetry.
Interaction records: Some first-party interaction events may include player prompt text or AI response text when needed to operate, evaluate, debug, or improve Wilds features.

How We Use Data

We use data to:

operate the product and generate AI responses
save your worlds, sessions, companions, and account state
measure usage, reliability, retention, payout attribution, and creator analytics
detect abuse, investigate bugs, respond to support issues, and secure the platform
generate anonymized or aggregated benchmarks, statistics, and business reporting

How We Handle Your Content

Automated processing is required to run Wilds. Your content is processed by our application and the AI models you choose so Wilds can generate responses, memories, summaries, art, and other outputs.
We do not use advertising analytics to record your raw writing. We do not send your raw prompts, stories, or companion dialogue to Google Analytics, Clarity, or Plausible as analytics payloads.
First-party operational logs can include text. Some server-side interaction records may store prompt text or AI output text. These records are treated as restricted operational data, not public creator analytics.
Human access is limited. We do not routinely have employees read private content, but authorized personnel may access limited data when necessary for support, debugging, fraud/security investigations, abuse incidents, or legal compliance.
AI providers. When you use cloud AI models, relevant context is sent to the selected provider to generate outputs. Local-model mode can keep generation on your own infrastructure when configured.

Analytics and Cookies

Consent-Gated Client Analytics

Wilds uses a consent banner and settings controls for optional browser analytics.

PostHog is our default client analytics provider. It is loaded only when analytics consent is granted. We use it for manual pageviews, named product events, funnels, cohorts, and masked session recordings.
Legacy analytics providers such as Google Analytics 4, Microsoft Clarity, and self-hosted Plausible may be enabled in limited deployments, but they are not the default analytics stack.
Global Privacy Control. If your browser sends a Global Privacy Control signal, Wilds treats it as an override that keeps optional browser analytics off while that signal is active.

Session Replay Privacy

When PostHog session replay is enabled, Wilds configures it to mask text in replays. We also disable autocapture and send only named events to keep collection intentional and controlled.

First-Party Product Analytics

Wilds also records first-party server-side interaction events for product operation, creator attribution, engagement analytics, fraud review, and service improvement. These events are stored in our own application database rather than in third-party advertising trackers.

Signed-in users can disable new account-linked first-party product analytics in account settings. This account-level control is separate from the browser cookie banner.

Guests are handled more conservatively: until a guest browser-analytics decision is synced server-side, new guest-linked interaction analytics stay off by default.

Research and Data Licensing Opt-In

Signed-in users can separately choose whether Wilds may include their de-identified interaction records in internal research, model-improvement, and any future external dataset or data-licensing program.

This setting is off by default.
This program-level opt-in is separate from routine product analytics.
Where applicable, eligible de-identified records can include prompt text or AI-response text.
We do not treat this setting as permission to disclose identified personal data.
Current export controls hash actor/session/entity identifiers and scrub direct text identifiers such as emails, URLs, phone numbers, and name-like metadata before staff dataset export.

Personas and Conversation Data

Personas are saved player identities you create. Each persona can hold its own conversations, mood and relationship state, and memory with any companion. You can maintain multiple personas under one account.

We store: the persona's name, pronouns, backstory, and any appearance fields you enter; per-persona companion chat messages; per-persona relationship and mood state tracked by companions; per-persona uploaded attachments; and (on Plus or higher) AI-generated persona avatars.

Deleting a persona

When you delete a persona you choose one of two modes:

Keep conversations. The persona record is removed and disconnected from past chats. Those chats remain readable and exportable as "Archived conversations."
Erase everything. The persona and every companion conversation, memory trace, conversation branch, and attachment that was scoped to it are permanently deleted from our primary stores, including the per-persona memory file on disk.

Account deletion always erases every persona and every persona-scoped conversation regardless of which deletion mode you used earlier. See the "Deletion and Retention" section below.

Exporting your persona data

The account export includes each persona's metadata (name, pronouns, backstory, appearance, avatar URL) and counts of conversations per persona with first/last activity timestamps, so you can see what you have before deleting.

Deletion and Retention

Content deletion: When you delete worlds, sessions, companions, personas, or similar user content, we permanently remove that content from the primary application data store, subject to legal holds and technical backup windows.
Account deletion: When you delete your account, core account content (including every persona and all persona-scoped conversation data) is deleted and the account shell may remain soft-deleted for a limited restore window. We also delete first-party interaction analytics linked to that account where technically keyed by account ID. Anonymous aggregate statistics may remain.
Billing records: Billing and tax records may be retained for as long as required by law.
De-identified reporting: We may retain anonymized or aggregated statistics that no longer identify you.

We Do Not Sell Raw User Content

Wilds does not sell your raw prompts, full transcripts, or companion conversations as standalone personal data products.

We may use or disclose:

aggregated or de-identified engagement metrics
world-level and creator-level public statistics
internal benchmarking and diligence reporting

Wilds now exposes a separate in-product research and data-licensing opt-in for signed-in users. Any future program involving text-bearing interaction records must honor that explicit setting and applicable law.

Your Choices

Cookie preferences: Use the consent banner, footer "Cookie Settings" link, or Settings > Tracking & Consent to control optional browser analytics cookies/storage.
Global Privacy Control: If your browser or extension enables GPC, Wilds keeps optional browser analytics off while that signal is present.
Guest analytics: Guest first-party interaction analytics follow the browser analytics choice once that consent is synced to the server.
First-party product analytics: Signed-in users can enable or disable new account-linked product analytics in Settings > Privacy, Stats & Data Use.
Research and data licensing: Signed-in users can separately opt in or opt out in Settings > Privacy, Stats & Data Use.
Account data: You can request export or deletion through account settings or by contacting team@wilds.ai.
Support and privacy requests: Contact team@wilds.ai.

California Residents

The California Consumer Privacy Act (CCPA), as amended, gives California residents rights including the right to know, delete, correct, and opt out of the sale or sharing of personal information, plus the right to receive notice at or before collection. We describe the categories of information we collect and the purposes for which we use them in this policy.

Wilds does not sell personal information in the ordinary sense of data-broker resale, and we do not sell raw prompts or transcripts. To exercise CCPA rights, contact team@wilds.ai.

Nevada Residents

Nevada residents have the right to opt out of the sale of certain covered information. Wilds does not sell covered information for third-party marketing purposes. Requests can be sent to team@wilds.ai.

Data Security

Passwords are hashed before storage.
Authentication uses signed tokens and server-side validation.
Production database and object storage access are restricted to authorized application and operations workflows.
Analytics tools are configured to minimize unnecessary capture and to respect consent controls.
PostHog session replay is configured with masked text, and research/data-licensing eligibility stays behind a separate user opt-in.
Staff-only interaction dataset exports require a server-side export salt and de-identification pass before records leave the primary analytics store.

Children

Wilds.ai is not directed to children under 18. If you believe a minor created an account, contact team@wilds.ai and we will investigate and remove the account if appropriate.

Service Provider

wilds.ai is operated by Manic Agency LLC, 2831 St Rose Parkway, Suite 200, Henderson, NV 89052.

Contact

For privacy questions or to exercise your data rights, contact: team@wilds.ai