Privacy Policy

1. What this policy covers

This policy covers data processed when you create an account, chat with an AI model, use search or code features, share a conversation, view history, or manage a subscription. It distinguishes content stored in your account, data used to measure usage, and temporary logs used to operate and secure WildGPT.

2. What WildGPT keeps in your account

We store the account information needed to operate the product: user identifier, email address, display name, username when present, avatar, language or preferences, current plan, creation and update dates, and the version and date of Terms acceptance.

A standard conversation stores its title, prompts, responses, selected model and, when produced by a feature, displayed reasoning, searches, or related tool activity. This content restores history and lets you continue a conversation across devices.

Sharing links have their own identifiers and permissions. Anyone with the appropriate access may view shared content. Remove or disable the share if you no longer want it available.

Stripe processes purchases. WildGPT stores customer and subscription identifiers, status, purchased plan, and events needed to track the subscription. We do not receive or store your full card number.

Do not enter passwords, API keys, health data, financial information, business secrets, or third-party data you are not allowed to submit.

3. What is sent to AI models

When you send a message, WildGPT sends the model provider the instructions and context needed to generate a response. A search query, code, or other relevant material may also be processed depending on the requested feature.

Providers act under their own commitments and may process data in other countries. WildGPT does not sell your conversations to model providers; the transfer is made to perform the request you initiated.

4. Usage, logs, temporary mode, and deletion

We count requests, estimated or billed tokens, usage reservations, credits, time-window limits, concurrent requests, and some attempts protected by anti-abuse controls. This data displays usage, enforces your plan, controls cost, and keeps the service available.

Analytics events may include the model, plan, view type, whether a conversation is new or temporary, and a prompt-length bucket. They are not designed to include the prompt text.

For a non-temporary conversation, WildGPT creates a temporary log for the sent message and received response. It includes message role and length, model, date, useful account or conversation identifiers and, when known, token count.

Message text may also be stored when content logging is enabled. In that case, only the first 4,000 characters are kept, and an automatic filter attempts to redact common passwords, tokens, and keys. The filter is not infallible, so never rely on it to protect a secret.

These logs automatically expire after 14 days. They support diagnostics, security, error investigation, and abuse prevention.

Temporary mode does not synchronize the conversation into history and does not send its messages to the application logs described above. Counters needed for quotas and cost may still be updated.

The request and context still have to be processed in memory and sent to the model provider to obtain a response. Providers may also maintain their own network or security logs under their policies and obligations, so temporary mode is not a promise of zero technical processing.

Standard conversations remain in your history until deletion. Deleting one removes its primary content and shares. A minimal record including its identifier and deletion date may remain to stop an old local copy from recreating it during synchronization.

An existing application log may remain until its 14-day period ends. Account, subscription, consent, usage, and security data is kept as needed to operate the account, bill correctly, resolve disputes, prevent abuse, or meet legal obligations. Technical copies or provider-held data may need additional time to disappear: deletion in the interface is not presented as instant erasure of every copy throughout the infrastructure.

5. Cookies, local storage, and security

WildGPT uses cookies and browser storage to maintain sessions, remember preferences, keep some local state, and protect account access. Blocking them may prevent authentication or break features.

We use reasonable access controls, permission separation, and monitoring. No online service is perfectly secure; protect your credentials and report suspicious activity quickly.

6. Providers and transfers

We use providers for authentication, hosting, storage, AI models, analytics, security, and payments. They receive the data needed for their function, subject to their own terms and applicable law.

Processing may occur outside your country. Where required, transfers rely on an appropriate legal mechanism or safeguard.

7. Your rights

Depending on applicable law, you may request access, correction, deletion, export, or restriction of certain data, or object to certain processing. We may verify your identity and keep information that the law requires or permits us to retain.

You can delete individual conversations in WildGPT. For a broader account or data request, use the support or contact channel provided in the product.

8. Minors, changes, and contact

WildGPT is not intended for anyone who is not old enough to accept these practices on their own under applicable law. This policy may change with the product or law, and we may provide additional notice for a material change. For privacy questions or requests, contact the WildGPT operator through the support or contact channel provided in the product.