Pudding
Product Privacy Support
中文Download ↗

Your data, explained

Privacy Policy

This policy explains how the Pudding desktop app, official website, and supporting authorization services at teatak.com handle your information.

Last updated: October 27, 2026

1. Scope

Pudding is developed and operated by Pudding Team. In this policy, “we” refers to Pudding Team. For privacy questions, contact yangglive@gmail.com.

This policy applies to the Pudding desktop app, official website, and supporting authorization services. Pudding itself does not require an account. Connecting a model or third-party service may require an account or credentials for that service.

Third-party models, plugins, websites, and services also have their own privacy policies. This policy describes Pudding’s practices.

2. Information we process and how we use it

Pudding stores conversations, attachments, generated content, settings, project references, Studio documents and tables, plugin configurations, and connection credentials on your device. Project files remain in the directories you choose.

When carrying out tasks, Pudding processes files, webpages, and third-party tool results you provide or authorize access to. It may also process screen or camera images and voice input when you use those features. Availability varies by operating system and version.

When connecting a service, Pudding processes information identifying the account, API keys, OAuth authorization codes, access tokens, or refresh tokens to establish and maintain your authorized connection.

Hosting and security providers for the website and authorization services may process IP addresses, request times, client information, and requested URLs to deliver services, troubleshoot problems, and provide security protection. When you contact us, we receive the contact details and issue information you provide.

3. Model training and third-party processing

We do not collect user content for model training or use user content to train models. Local storage, task processing, service connections and support needed to provide features remain subject to the practices described in this policy.

When you use a third-party model, the input, context and tool results needed to complete your request may be sent to the model service or API intermediary you configure. Whether the recipient retains the data, permits human access or uses it for training depends on the specific service terms and account configuration. Using your own API key does not automatically mean that data will not be retained or used for training. Google user data is subject to the additional restrictions in section 5.

If a model runs on your device and the task does not call external services, the relevant model processing can take place locally. Choosing a local model does not automatically disable the browser, plugins, authorization refreshes, or other network features.

Third-party services may also process information outside your country or region.

4. Gmail and other connections

Connecting Gmail is optional. The integration supporting reading and sending uses basic Google account information to identify the connected account, gmail.readonly to search and read emails, and gmail.send to send emails on your behalf. The specific permissions requested are displayed during authorization.

Data processed may include senders, recipients, subjects, message bodies, attachments, and message identifiers to complete your requested email tasks. These permissions do not allow email deletion or Gmail draft management.

When you ask a model to process emails, relevant email content or derived context may be included in the model request. When sending an email, its content and attachments are transmitted to Google and delivered to the recipients you specify.

For GitHub and other plugins, the data processed and actions available depend on the plugin’s features, granted permissions, and your task. Examples include account information, repository content, issues, and pull requests.

5. Restrictions on Google user data

Use and transfer of information received through Google APIs, including derived data, are subject to the Google API Services User Data Policy and Google Workspace user data and developer policy, including the Limited Use requirements. This data is used only for user-facing features you authorize and that are clearly presented in Pudding.

We do not sell personal information or use Google user data for advertising, data brokerage, or credit decisions. User data obtained through Google APIs, including derived data, must not be used to develop, train or improve AI or machine learning models for other users or shared use. This restriction also applies to data transferred to third-party model services or API intermediaries. It is not waived by a user choosing the service, providing their own API key or consenting to the transfer.

Transfers of Google data to third parties are limited to what is needed, with your consent, for relevant user-facing features, or for security or applicable legal requirements. Human access is limited to circumstances permitted by Google’s policies, such as explicit consent to view specific data, necessary security investigations, or applicable legal requirements.

These restrictions also apply to model services or intermediaries receiving the data. Third-party terms or user consent do not waive them. This policy does not represent that Pudding has verified the data-use rules and practices of every service you configure. Providing your own API key does not, by itself, establish that these restrictions are met.

6. Authorization services and recipients

Pudding’s OAuth services use teatak.com to process authorization codes and tokens needed for connections. Gmail authorization codes and refresh tokens are transmitted to this service for exchange or refresh; these endpoints do not write authorization codes or tokens to persistent application storage. GitHub authorization results are temporarily stored on the server for retrieval by the app. Connection credentials are also stored on your device.

Cloudflare provides hosting or network services for the website and supporting authorization services. Google, GitHub, and the model or plugin services you configure receive information needed to complete the corresponding requests. We may also disclose information as required by applicable law or for necessary security investigations. Google user data remains subject to section 5.

7. Retention, deletion, and revoking access

Local conversations, files, settings, and credentials may remain until you delete them through the app, a file manager, or your operating system. Deleting original content does not necessarily delete references in conversations, generated results, exports, synchronized copies, or backups.

GitHub authorization results are available for retrieval for five minutes and deleted after successful retrieval. Unclaimed records are cleared after expiry. The statement that Gmail endpoints do not write to persistent application storage does not mean hosting or security services do not process request logs.

The official website and OAuth authorization services run on the same Cloudflare Worker. Workers Logs and Workers Traces are currently disabled for this Worker. Cloudflare may still process other request, security, or infrastructure information; these disabled settings do not mean that historical logs have been deleted.

Support emails and related correspondence are retained indefinitely; no automatic deletion period is currently set.

Removing a connection, revoking the provider’s authorization, and deleting local data are separate actions. Some connections attempt to revoke authorization when removed. You can also review and revoke access in the provider’s account settings. Revoking access does not automatically delete local history, generated content, or backups, recall sent emails, or delete recipients’ copies.

For help accessing or deleting information under our control, contact us using the email below. You can also consult the data deletion guide.

8. Website storage and security

The website initially follows your browser’s language preferences. When you manually change languages, the functional cookie teatak_lang remembers your Chinese or English choice for one year across teatak.com and its subdomains. The website also uses local storage named teatak.theme to save your theme preference. You can clear these records through your browser settings.

We use HTTPS to protect network transmissions for the website and authorization services we operate. No measure guarantees absolute security. Local data and credential security also depends on your device, operating-system account, and backup settings. If you suspect a leak, promptly revoke the relevant credentials or authorization and contact us.

9. Your choices and children’s information

You can choose not to connect optional services, adjust project and tool permissions, choose a different model, or stop relevant tasks. Permission changes do not automatically delete previously processed data.

Pudding is not directed to children under 13. If you believe a child has provided personal information to us, contact us so we can investigate and take appropriate action.

10. Updates and contact

We update this policy as the product and information-handling practices change. If the purposes for using Google user data change, we will explain the new purposes and obtain the required consent before beginning that use.

For privacy questions, data requests, or security reports, contact yangglive@gmail.com. Do not include passwords, API keys, or complete tokens.

Pudding

A desktop AI workspace with an open-source core.

ProductWhat it doesDownloadGitHub
Privacy & termsPrivacyTermsData deletion
HelpSupportContact
© 2026 PuddingDesigned for work that stays yours.