J.doe

J.doe Privacy Policy

Effective date: October 20, 2026
Last updated: October 20, 2026

J.doe is operated by J.doe Limited, a company registered in Hong Kong ("J.doe," "we," "us," or "our"). This Policy explains how we handle personal information through myjdoe.com, our applications, and our AI services (together, the "Services"). The Services are currently offered as a free beta for adults.

We are responsible for the processing described in this Policy. You can contact us at support@myjdoe.com.

1. Information we process

Information you provide. Information you give us when using the Services or contacting support, including the contents of your messages and the contact details you provide.

Account and notification information. Your email address and authentication information needed to sign in, including information you authorize Apple or Google to share when using Sign in with Apple or Sign in with Google. We also process device tokens and notification information needed to deliver push notifications.

Original Content. Text you enter, audio you upload or record, screen content you authorize us to access, and content obtained from accounts you authorize us to connect to. This information may concern you or other people.

Generated and derived information. Transcripts, text extracted from screenshots, summaries, reminders, memories, AI responses, and other results created to provide the features you use.

Speaker features. Numerical features extracted from recordings and stored to match the same speaker across recordings. Features may be saved for you and other speakers in your recordings. These features are distinct from the original audio. Depending on applicable law and how the feature operates, this processing may involve biometric data.

Campaign information. Campaign-source information included in UTM-tagged links, such as source, medium, and campaign labels, as described in Section 6.

Crash and error diagnostics. On iOS, we use Google Firebase Crashlytics to diagnose crashes and errors. Reports include device and operating-system information, crash traces, operation names, error types, code locations, and SDK-generated installation and session identifiers. We do not attach J.doe account IDs or your content to these reports.

2. Recording, screen access, and speaker matching

During registration, the required checkbox records your acceptance of the Terms, acknowledgment of this Policy, and authorization for screen capture, audio recording, and cloud AI processing described in the accompanying notice. We do not ask you to complete a second J.doe consent page when you first use screen or recording features within that authorization's scope. Registration does not itself grant operating-system permissions: actual access follows your device permissions and feature settings. Screen context can include screenshots and text visible in other applications. You may also upload recordings.

You can change the relevant permissions through the applicable device or app settings. Withdrawing microphone or screen permission stops future access under that permission but does not, by itself, delete information already provided. Turning off cross-recording speaker recognition deletes stored speaker features from the active service, as described in Section 9. A feature may be unavailable if you decline the information or permission it needs.

When turning off screen or recording access, you can also choose to delete the associated previously uploaded data. This option is unchecked by default. The information deleted and retained is explained in Section 9.

The registration authorization for screen capture and recording does not include authorization to store speaker features for cross-recording recognition. That feature has a separate, optional authorization covering features of you and other speakers in your recordings. The user consents to processing their own features and confirms that they have obtained any consent required from other speakers. This confirmation does not itself constitute another speaker's consent or remove J.doe's obligations under applicable law. Speaker matching uses stored numerical features to associate the same speaker across recordings. We retain the extracted features for matching and encrypt them separately. You can choose to delete these features. Their retention and destruction are described in Section 9.

Your device permission does not establish that everyone whose voice or information is captured has consented. Before providing another person's information, obtain any permissions required by law. Avoid capturing passwords, authentication codes, or sensitive material unnecessary for your request. J.doe remains responsible for its own obligations under applicable law.

3. AI processing, training, and human access

We use AI models through two cloud AI services, each under our own account: Amazon Bedrock (AWS) and, as an alternative for some features, Google Cloud Vertex AI. Our use of these services for AI processing is limited to the United States, as described in Section 8.

For these J.doe-operated features, we do not directly call OpenAI's or Anthropic's APIs or use third-party model-routing services. Models accessed through Amazon Bedrock are operated by AWS; third-party model developers, such as Anthropic and Cohere, do not receive your content through that processing. Gemini is processed by Google through Vertex AI. Google does not use this content to train or fine-tune its models without our permission, and we do not give that permission. User-directed handoffs to external tools are addressed separately in Section 4.

We do not use Original Content to train AI models. The J.doe team does not manually read, listen to, or review Original Content held within these AI features. These commitments apply to Original Content and describe J.doe's own handling of that content.

Amazon Bedrock retention. We configure Amazon Bedrock for zero data retention for the models and Regions we use. Under this configuration, AWS does not retain AI inputs and outputs for routine safety review or conduct human review for that purpose, and we do not use models that require such retention. This does not affect content stored by J.doe in its own AWS infrastructure under Section 9, or AWS's limited retention and review of flagged content to meet applicable legal reporting obligations.

Google safety review. When we use Gemini through Vertex AI, Google may retain prompts flagged for further investigation for up to 90 days for abuse monitoring, and authorized Google personnel may review them. These Google logs are stored in the U.S. region or multi-region selected for our project. These safety-review copies are not used to train models. Any retention required by law is separate from this safety-review period. Provider safety review is separate from the commitments concerning J.doe's own training and human access above.

Information you choose to send separately to our support email is read to respond to you. Sending a support request does not itself authorize staff to access your content inside the Services.

4. Connected accounts and tools you choose

Connected accounts. You may authorize connectors for services such as Gmail, Slack, Outlook, Notion, or Lark. We read information within the permissions you grant and use it to provide your requested features. Authorizing a read connection does not, by itself, authorize sending messages, changing records, or deleting source information.

You can revoke the relevant authorization through the connected service's permission controls. Revocation stops further access under that authorization but does not automatically delete copies already received by J.doe.

Google API data. J.doe follows the Google API Services User Data Policy, including its Limited Use requirements, when using or sharing information obtained through Google APIs and data derived from that information.

Your own coding tools. If you instruct J.doe to hand a task to Claude Code or Codex installed on your computer and signed in to your own account, task content and relevant files may be processed by that tool and its provider under your agreement with them. This processing uses your account rather than J.doe's AWS account. Their privacy, training, retention, and location terms may differ from ours. J.doe's commitments continue to apply to information we retain or process ourselves.

Web search. We use AWS AgentCore Web Search. When you use web search, we send only AI-generated search terms based on your question, excluding passwords, verification codes, payment-card numbers, and identity-document numbers. Search terms are processed by AWS in the United States and are not sent to third-party search engines. AWS personnel do not have access to these search terms. You can turn web search off in settings.

When retrieving a web page for you, the request comes from J.doe's servers rather than your device. The target website receives that server request. J.doe's page access is read-only: it does not sign in or submit forms and respects the site's robots.txt rules. Details for website operators are available on our web-access page.

5. How we use information

We use information to provide the AI features you request, recognize speakers across recordings, operate authorized connections, process your deletion choices and requests for a copy of your information, respond to support requests, and understand the campaign sources of visits to our website. We also use information to authenticate sign-ins, send verification-code emails, deliver push notifications, and diagnose crashes and errors. We process information where necessary to meet applicable legal obligations.

If speaker matching or other processing involves special-category personal data, an additional processing condition is required under applicable law. Permission from an account holder does not replace any separate consent required from another person whose data is processed.

You do not have to provide optional information. Where we rely on consent, you may withdraw it without affecting the lawfulness of earlier processing.

We use UTM-tagged links to understand which campaigns or sources bring visitors to our website. UTM parameters are labels included in a link, such as a source name, medium, or campaign name. We use this campaign information to understand the effectiveness of our outreach.

7. When information is disclosed

AWS. We use AWS to host and process information for our core service, including AI processing through Amazon Bedrock and website delivery through Amazon CloudFront. Our Bedrock zero-retention configuration and limited legal-reporting exceptions are described in Section 3.

AWS also processes generated search terms for the web-search feature described in Section 4. Website operators receive requests from J.doe's servers when pages are retrieved.

Google Cloud. For some AI features, we use Google Cloud Vertex AI to process the content needed for the feature, as described in Section 3.

Google Firebase Crashlytics. We use Crashlytics for crash and error diagnostics on iOS. Google processes the diagnostic information described in Section 1 to help us identify and resolve technical problems.

Supabase. We use Supabase for account authentication. Supabase processes your email address and authentication information needed to provide sign-in functionality.

Google Workspace. We use Google Workspace to send verification-code emails. This service processes your email address and the verification messages needed for delivery.

Apple. When you choose Sign in with Apple, Apple processes information needed to authenticate your sign-in. We also use Apple Push Notification service (APNs), which processes device tokens and notification information needed to deliver push notifications.

Google. When you choose Sign in with Google, Google processes information needed to authenticate your sign-in.

At your direction. Information may be exchanged as needed for connections or tasks you authorize, as described in Section 4.

Legal requirements. We may disclose information where required by applicable law or a valid legal process, limited to what is appropriate to that requirement.

Business changes. Information may be transferred in connection with a merger, acquisition, or transfer of the Services, subject to applicable law and required notices.

These disclosures do not grant us additional permission to train on or manually review Original Content beyond Section 3.

We maintain a service-provider list, including the cloud services used for AI processing. Before we add a new AI model provider, we will update that list and give you advance notice through the Services or by email.

8. Storage, location, and security

Our service-content storage and AI processing take place in the United States. Your content is stored in the AWS US East (Northern Virginia) Region. AI requests may be processed in other AWS Regions within the United States, or in Google Cloud regions within the United States when we use Vertex AI, depending on the model and service configuration. We do not use configurations that route these J.doe-operated AI requests outside the United States. Google safety-review copies described in Section 3 are stored in the United States. Account authentication data held by Supabase is also stored in the United States. Data stored in our AWS environment is encrypted. Our operator is registered in Hong Kong, and your information may be processed outside the country where you live.

If you are located in the European Economic Area, the United Kingdom, or another jurisdiction that restricts international transfers, your information is transferred to the United States and may be accessed from other countries where our team operates, subject to the restrictions on J.doe's access in Section 3. Where required, we rely on transfer safeguards included in the applicable service-provider agreements, such as the European Commission's Standard Contractual Clauses and the UK Addendum where applicable. You can contact support@myjdoe.com for more information about these safeguards.

These U.S. location statements cover service-content storage, J.doe-operated AI processing, and the Supabase account data described above. Our public website is delivered through Amazon CloudFront's global network. Google Workspace, Google (for sign-in), Apple, and Firebase Crashlytics provide the email, sign-in, notification, and diagnostic services described in Section 7; their processing locations may differ from the United States. Independent services you authorize, such as your own Claude Code or Codex account, also have their own processing locations and terms. You can contact support@myjdoe.com with questions about international processing.

No online service can guarantee absolute security. Please report concerns to support@myjdoe.com. We will provide breach notices required by applicable law.

9. Retention and deletion

Service content. The following are the default retention periods for the listed data, subject to the exceptions below and your choice to delete it earlier:

Card evidence. If a screenshot, extracted text, transcript, or audio clip is used as the basis for a card, it is retained for up to 30 days after that card is created. An audio clip is deleted earlier if its associated card becomes invalid before that period ends. This rule concerns the clip used as evidence, not an extension of the default retention period for the entire original recording.

Items marked to keep. If you mark an item to keep, it is retained until you remove that mark or delete the item. Removing the mark ends this retention exception. Items marked to keep are still deleted when they fall within your request to delete previously uploaded data, or when you delete your account. Stored speaker features follow the separate rule below.

Other generated information, such as memories, summaries, and reminders, is retained as needed to provide the relevant feature and is covered by your deletion choices below.

Web-search records. We do not store generated search terms or retrieved webpage bodies in our application database. We retain only cited URLs and page titles as search-reference records with the conversation and delete them with that conversation.

Deleting data when turning off screen or recording access. Turning off access stops future collection under that permission. If you also select the option to delete previously uploaded data:

For both options, content within that scope is deleted even if you previously marked it to keep. We provide a deletion receipt when the deletion is complete. These options do not delete new data collected after you re-enable access, stored speaker features, or existing cards and reminders. Cards and reminders remain, but their references to deleted screenshots or recordings are shown as deleted. Speaker features have their own control, described below.

If you do not select deletion, the previously uploaded data remains subject to its applicable retention periods, and memories already derived from it can continue to be used. These feature-specific choices are separate from requesting deletion of all user content or your account.

Speaker features. We do not sell stored speaker features. We automatically delete the features saved in your account after three consecutive months without use of the cross-recording speaker-recognition feature on your account. You can also delete them at any time. Turning off this feature immediately deletes all saved speaker features from the active service, with no option to keep them while the feature is off. Account deletion also deletes these features. Database backup copies expire as described below.

Deletion requests. You can request deletion of all your J.doe user content, including Original Content and associated transcripts, screen-derived information, summaries, memories, reminders, and stored speaker features. You can also contact support@myjdoe.com to request deletion of personal information or your account. Once an account-deletion request is submitted, it cannot be withdrawn: the account is immediately frozen, and cleanup continues in the background. This does not mean that every copy is erased immediately.

Limited records after account deletion. We remove direct identifying information and retain only the minimum records needed for the following purposes:

These are limited administrative records, not permission to retain your recordings, screenshots, conversations, or other user content for those periods. Database backup copies expire within seven days. Any additional retention required by law is limited to that requirement.

Provider retention. Prompts retained by Google through Vertex AI for safety and abuse prevention may remain after you delete your J.doe data for the remainder of the original safety-review period of up to 90 days, unless retention is required by law. Those copies remain subject to Section 3. For Firebase Crashlytics, Google retains diagnostic data and associated identifiers for 90 days before beginning deletion from its live and backup systems. Bedrock processing is subject to the zero-retention configuration and limited legal-reporting exceptions described in Section 3.

Deleting J.doe data does not delete the original messages or files in connected accounts, or information independently held by a tool you asked to handle a task. Those services provide their own deletion controls. Uninstalling J.doe or revoking a device permission does not itself delete stored J.doe data.

10. Your privacy rights

Depending on your location and applicable law, you may have rights to access, correct, delete, or receive a portable copy of your personal information; restrict processing; withdraw consent; or object to certain uses. Rights may be subject to legal conditions and exceptions.

Copy of your information. To receive a copy of your account information, profile information, consent records, device list, text from collection records, conversation history, and audio transcripts, email support@myjdoe.com. We will verify your request and provide the copy in a commonly used electronic format within the time required by applicable law. Original screenshots and original audio files are kept only for the periods described in Section 9. This does not limit any broader access or portability rights you have under applicable law.

Right to object. Where EEA or UK law applies, you may object to processing based on legitimate interests because of your particular situation. You may object at any time to processing for direct marketing.

Email support@myjdoe.com to exercise a right. We may request proportionate information to verify your identity or an authorized agent's authority. We will respond within the time required by applicable law and explain any lawful reason for refusing a request. Where an appeal right applies, you may request review by replying to our response or emailing the same address. We will not unlawfully discriminate against you for exercising your rights.

You may also contact us if your information appears in content another user provides. Include enough context to help locate it without sending unnecessary sensitive details.

You may complain to a competent authority. EEA residents can find their authority through the European Data Protection Board; UK residents can contact the ICO. You may also contact the Hong Kong Privacy Commissioner for Personal Data.

11. Adults only

The Services are intended only for people who are at least 18 and have reached the age of majority where they live. Minors may not use the Services. If you believe a minor has created an account or provided information directly to us, contact support@myjdoe.com so we can investigate and take appropriate action.

12. Changes and contact

We may update this Policy as the Services or legal requirements change. We will revise the date above and provide appropriate additional notice of material changes. We will obtain consent where required; changing this Policy does not itself authorize a new data use requiring consent.

Operator: J.doe Limited
Registered address: Room 704, 7/F., Tower A, New Mandarin Plaza, 14 Science Museum Road, TST East, Kowloon, Hong Kong
Email: support@myjdoe.com
Website: myjdoe.com