DHwork Privacy Protection Guidelines
Version:
V1.0Release Date:July 13, 2026Effective Date:July 17, 2026
Important Notice
These Guidelines apply to the AI/agent-related features provided by DHwork and cover scenarios including account and enterprise management, intelligent conversations and automated operations, local terminal permissions, third-party models/tools, enterprise data, cookies, sharing and entrusted processing, storage security, user rights, and protection of minors. The specific scope of processing depends on the features you use, your organization’s configuration, and the scope of your authorization.
DHwork is an AI/agent product that can invoke artificial intelligence models, software tools, workflows, local devices, cloud resources, and third-party services within the scope authorized by users. An agent may read content, generate results, or initiate operations according to your instructions; however, the features available vary across versions, devices, organizational configurations, and the third-party services you connect.
If a feature is not enabled, you have not granted the necessary permissions, or your version does not support it, DHwork will not process the corresponding information on account of that feature. We will explain the specific processing activities through feature pages, system permission prompts, authorization pages, explanations in the administration console, third-party sharing lists, or supplementary rules; these explanations form part of these Guidelines to the extent applicable.
I. Definitions and Scope
1.1 Definitions
Unless the context provides otherwise:
DHwork / the Service / we / us: The DHwork clients, web version, mobile version, administration console, APIs, cloud capabilities, and related support services operated by 数贸科技（北京）有限公司.
User / you: A natural person using the Service; if you use the Service on behalf of an organization, this also includes administrators and members authorized by that organization.
Organizational Customer: An entity that activates, purchases, or manages the Service in the name of an enterprise, public institution, social organization, government agency, or other organization.
Administrator: A person designated by an Organizational Customer who can configure the organization, members, permissions, data sources, models, tools, auditing, or security policies.
Organizational Member: An employee, contracted worker, partner, or other authorized person who uses the Service within the management scope of an Organizational Customer.
Personal Information: Any information recorded electronically or by other means that relates to an identified or identifiable natural person, excluding anonymized information.
Sensitive Personal Information: Personal Information that, once leaked or unlawfully used, is likely to impair a natural person’s dignity or endanger their personal or property safety, such as biometric information, religious beliefs, specific identity status, medical and health information, financial accounts, whereabouts and movement records, and the Personal Information of minors under the age of fourteen.
Input Content: Text, images, audio and video, files, code, web page content, instructions, parameters, credentials, or other content that you actively submit, upload, paste, enter by voice, authorize to be read, or that a workflow obtains within the scope of your authorization.
Output Content: Text, code, files, page operation results, task records, or other content that the Service, or a model, tool, or workflow you select, generates, retrieves, converts, summarizes, executes, or returns based on Input Content.
Tools/Connectors: Local tools, browser automation, plugins, skills, data sources, APIs, cloud services, or third-party application integrations that you enable or configure to complete tasks.
Enterprise Data (Organizational Data): Data submitted, generated, configured, or controlled by an Organizational Customer or Organizational Member within an organizational account, on an organizational device, in an organizational space, or within the scope authorized by the organization, including organizational structures, accounts and permissions, knowledge bases, files, conversations, tasks, auditing, and resource usage data. Enterprise Data may contain Personal Information.
1.2 Scope
These Guidelines explain our rules for processing Personal Information when providing DHwork, including collection, storage, use, processing, transmission, provision, disclosure, and deletion, and apply to PC, Web, mobile, APIs, remote collaboration, and the administration console.
These Guidelines do not automatically apply to independent third-party products or services that you select, configure, or access through a redirect yourself (including external large models, plugins, websites, cloud drives, payment, identity authentication, or communication services). Such third parties will process information under their own rules. If we engage a service provider to process Personal Information for a DHwork feature, or jointly determine the purposes and means of processing with a third party, we will separately explain the relevant relationship, processing scope, and responsibilities as required by law.
II. Principles, Grounds, and Sources for Our Processing of Personal Information
2.1 Basic Principles
We process Personal Information in accordance with the principles of lawfulness, legitimacy, necessity, good faith, openness, and transparency, and take security measures appropriate to the processing purposes. We will follow data minimization: we process only the information needed to implement a specific feature; refusal to authorize a non-essential feature will not affect basic services (unless laws or regulations provide otherwise or the information is necessary for that feature).
2.2 Grounds for Processing
Under applicable law, we will process Personal Information where consent has been obtained, where necessary to enter into or perform a contract with you, to fulfill statutory obligations, to respond to emergencies, to process information you have made public yourself or other lawfully public information, or in other circumstances with a statutory basis. For matters that require separate consent by law, such as processing Sensitive Personal Information, providing Personal Information to other Personal Information processors, or providing Personal Information overseas, we will obtain the corresponding consent or satisfy other applicable statutory conditions as required by law.
2.3 Sources of Information
We may obtain information through the following channels:
Information actively entered, uploaded, configured, authorized, or submitted by you, your Administrator, or an authorized person;
Service, device, network, operation, and security logs automatically generated when you use the Service;
Third-party accounts, data sources, Connectors, models, or tools authorized by you;
Information provided by an Organizational Customer to activate, manage, and safeguard organizational services;
Service-related information obtained from partners or public sources to the extent permitted by law.
Please do not submit another person’s Personal Information, trade secrets, or content subject to confidentiality obligations to us without a lawful basis. If you act on behalf of an organization, create accounts for others, upload address books/data libraries, or authorize DHwork to access third-party systems, you should ensure that the necessary notices have been given, valid authorization has been obtained, or another lawful basis for processing exists.
III. How We Process Information: Features, Categories, Purposes, and Necessity
The following table describes common processing scenarios. The actual processing scope depends on the features you enable, the scope of your authorization, and the notices on the relevant pages; “optional” means that refusal may prevent you from using the corresponding extended feature, but should not affect unrelated basic features.
| Feature/Scenario | Information That May Be Processed | Purpose and Necessity | Whether Optional and Consequences of Refusal |
|---|---|---|---|
| Registration, login, and account security | Account identifier, nickname, avatar, mobile number/email address, verification information for passwords or authentication credentials, login time, login IP, device and session identifiers | Create, identify, and protect accounts, prevent unauthorized access, and provide account recovery and notification services | Account information is usually necessary for activation or login; the authentication method is as shown on the relevant page |
| Ordering, organizational activation, and administration | Organization name, Administrator/Member names or nicknames, employee numbers, departments, positions, organizational relationships, roles and permissions, order or authorization status | Configure organizational spaces, member collaboration, seats/permissions, and service support | Necessary for organizational features; Organizational Members may consult their Administrator about the organization’s rules |
| Intelligent conversations, task orchestration, and result generation | Input Content, attachments and their metadata, conversation context, prompts, task parameters, Output Content, task status and times | Understand requests, invoke models/tools you select or that the platform provides, generate or deliver results, and maintain conversation continuity | Usually necessary for the task; you may choose not to submit particular content or delete conversations (subject to feature support) |
| Files, data libraries, and knowledge bases | Contents of files you select, file names, paths/sources, formats, creation/modification times, indexes, and retrieved excerpts | Parse, retrieve, summarize, convert, answer questions about, or process files according to your instructions | Optional; corresponding file tasks cannot be completed without authorization or uploaded files |
| Local tools, command execution, or code execution | Commands, scripts, execution parameters, standard output/error, task artifacts, and necessary runtime status that you specify or that fall within the authorized scope | Complete automation, development, or processing tasks you explicitly request on a local device | Optional; if you refuse, you will need to execute the task locally yourself. A confirmation mechanism should be provided for high-risk operations . |
| Browser control or third-party software automation | Page content, form fields, operation steps, screenshots/page status, and execution results to which you grant access; these may include information in third-party account sessions | Complete queries, data entry, testing, or process operations on your behalf | Optional; automation is unavailable without authorization. Data processing on third-party pages is also subject to third-party rules |
| Voice, image, or screen-based input (if enabled) | Microphone audio, voice transcription text, images/screenshots/screen content you select, and metadata | Receive multimodal instructions and recognize or generate content | Optional; refusal of the corresponding system permission does not affect other features that do not depend on that permission |
| Memory, preferences, or personalization (if enabled) | Preferences, roles, project background, frequently used settings, and memory entries you provide or confirm in conversations | Provide more coherent and personalized responses in subsequent conversations | Optional; no new related memories will be added after this feature is turned off |
| Remote access, task synchronization, or cross-device collaboration (if enabled) | Linked account identifiers, device identifiers, remote instructions, task status, and necessary conversation/result content | Securely route instructions to devices you have authorized or synchronize task status | Optional; remote capabilities are unavailable if devices are not linked or the feature is not enabled |
| Customer service, support tickets, and feedback | Contact details, account information, problem descriptions, attachments, and communication and handling records | Respond to inquiries, identify problems, handle complaints, and improve the Service | You may choose not to provide non-essential information; insufficient necessary information may affect the handling of your issue |
| Security, stability, and compliance | Device model/system version, client version, network information, IP address, crash/performance logs, operational audit records, and anomaly or risk signals | Diagnose faults, prevent fraud and abuse, protect account and system security, and fulfill statutory obligations | Necessary for secure operation; we should limit log granularity and retention periods |
| Service improvement and statistics | Email address, name (if available), application version, platform, and feature usage and feedback records; as well as de-identified or aggregated performance, error, configuration, and resource consumption information | Improve compatibility, reliability, interaction, and capacity management through analytics services such as PostHog | The mobile version reports data after you agree to the data usage notice; usage statistics do not include message bodies or file contents |
| Notifications and marketing (where applicable) | Contact details, subscription preferences, and necessary service-related records | Send account, security, service, and order notifications; send marketing information after obtaining the necessary consent | Necessary service notifications may not be possible to disable; marketing information should offer a convenient way to unsubscribe |
3.1 Device, Network, and Log Information
To ensure normal operation of the client and account security, we may collect necessary device and network information, such as device type, operating system and version, application version, language/time zone, device or installation identifier, network type, IP address, log time, crash stack traces, and performance and error information. We should not use permanent identifiers that can directly identify a device instead of the smaller amount of data needed to achieve the same purpose; if a device identifier is genuinely necessary, its name, purpose, processing method, and retention period should be specified in the product list.
Logs may contain the context of your operations, and may even indirectly contain input excerpts because of error messages or request parameters. We should implement data masking, access controls, and the shortest retention periods, and prohibit the use of logs for purposes unrelated to security, troubleshooting, or authorized improvements.
3.2 Device Permissions, Scope of Access, and Withdrawal of Authorization
DHwork requests and uses necessary permissions only after you actively use the corresponding capability and complete system-level or in-product authorization. Granting permission does not mean continuous, unrestricted reading; we should process information according to the principles of “task-triggered, visible scope, minimal access, and stopping after use.”
| Permission/Capability (Where Applicable) | Purpose | Recommended Authorization and Control Requirements |
|---|---|---|
| Network access | Login, model or cloud service calls, synchronization, and updates | Explain the purpose of connecting to the network and the recipients; clarify offline capabilities where available |
| File reading | Read local or cloud files you explicitly select | Access only files you select or files within authorized directories; display the scope of files |
| File creation, writing, modification, or deletion | Save results, edit files, and clean up task artifacts | Provide confirmation, previews, or rollback capabilities for high-risk operations such as overwriting, bulk modification, or deletion |
| Clipboard | Use content you actively paste as input | Limit access to user pasting/explicit triggering; do not continuously read in the background |
| Microphone | Receive voice instructions and transcribe them | Display a clear status before recording; whether audio is uploaded or stored and its retention period must be confirmed |
| Screen recording, screenshots, or accessibility | Recognize screen content selected by the user or assist in completing operations | Explain the scope of collection before an operation; avoid collecting unrelated window/notification content |
| System commands and code execution | Execute local tasks according to instructions | Require prominent notice of the execution content, working directory, network access, and potential effects; restrict dangerous commands by default |
| Browser control | Access websites, read pages, fill in forms, or automate processes | Clearly identify target sites and the scope of operations; do not bypass third-party security verification or access controls |
You may disable optional permissions already granted through operating system settings or the DHwork settings page. After withdrawal, we will stop further processing information based on that permission; however, withdrawal does not affect processing already completed under valid authorization before withdrawal, and may make related features unavailable.
IV. Special Explanations Regarding Input, Output, Tool Calls, and Enterprise Data
4.1 Your Control over Input and Instructions
You decide whether to submit Input Content, enable tools, link data sources, or confirm an operation. Before entering content, please assess for yourself whether it contains Personal Information, Sensitive Personal Information, state secrets, trade secrets, information protected by confidentiality obligations, access tokens, or other sensitive credentials. Unless a feature explicitly requires it and you understand the risks, submitting such content is not recommended.
DHwork may generate output, propose actions, or invoke tools within the scope of your authorization based on your input. Generative AI output may be inaccurate, incomplete, outdated, or unsuitable for particular circumstances; before using it for legal, medical, financial, personnel, production safety, decisions affecting significant rights and interests, or external publication purposes, you should independently verify it and retain necessary human decision-making. DHwork should not make decisions with a significant impact on individuals solely on the basis of automated processing results without appropriate human review and permission under the law.
4.2 Processing Boundaries for Local, Cloud, and External Services
Local execution. If a task can be completed locally, DHwork may process information within the scope of devices, directories, and tools you authorize.
DHwork cloud capabilities. If you choose to use cloud models, retrieval, storage, synchronization, or workflow capabilities provided by the platform, the input, context, task parameters, output, or logs necessary to complete the request may be transmitted to DHwork servers or entrusted service providers. Specific data flows, storage locations, model service providers, retention periods, and other details are governed by Sections VII and VIII and the relevant lists.
External models, tools, or Connectors you connect. When you configure an API Key, OAuth authorization, third-party account, or data source yourself, DHwork may send the request content necessary to complete a task to the corresponding third party according to your instructions. The third party’s processing of information it receives is governed by its privacy rules and your agreement with it. Before connection, we should display the recipient, scope of permissions, and categories of data that may be sent; credentials such as API Keys and access tokens should be protected using encryption or a secure credential vault, and we should clarify whether they are stored only locally/in the organizational environment or will be stored in the cloud.
Boundaries of autonomous operations. For high-risk actions involving sending emails/messages, submitting forms, placing orders and making payments, deleting or overwriting files, publishing content, modifying production environments, accessing restricted systems, or similar activities, DHwork should provide appropriate previews, confirmation, tiered permissions, logs, and mechanisms for reversal; the specific mechanisms depend on the notices and permission settings of the corresponding feature.
4.3 Enterprise Data and the Role of Organizational Customers
When an Organizational Customer activates, configures, or manages organizational accounts and determines the purposes and means of processing relating to Organizational Members, data libraries, workflows, auditing, or security policies, the Organizational Customer is usually the Personal Information processor for the relevant Personal Information in Enterprise Data; the DHwork operator provides processing services within the scope of the Organizational Customer’s instructions and the parties’ agreement, usually as an entrusted processor. The specific roles should be determined based on the actual product architecture, contracts, and applicable law.
In organizational management scenarios, Administrators may, according to the organization’s configuration, access or export member account information, organizational structures, permissions, conversation/task metadata, resource usage, and audit logs, including Input Content, Output Content, file contents, screen/command logs, and their precise permissions. Organizational Customers should clearly inform members of their management scope, audit policies, data purposes, retention periods, and channels for exercising rights, and ensure that processing has a lawful basis. If Organizational Members have questions about data controlled by their Organizational Customer or wish to exercise their rights, they may first contact the Organizational Customer or Administrator; we will assist to the extent permitted by law and contract.
V. Sensitive Personal Information
Certain DHwork features may process Sensitive Personal Information because of your active input, uploads, authorization of access, or organizational configuration, such as identity document information, biometric information including facial/voiceprint information, health information, financial account information, precise location, information about minors, authentication materials, or files, conversations, and output that can reveal the foregoing information.
We process Sensitive Personal Information only where there is a specific purpose, sufficient necessity, and strict protective measures, and will provide prominent notice, obtain separate consent, or satisfy other statutory processing conditions as required by law. The specific types of sensitive information, purposes and means of processing, retention periods, recipients, and effects on individuals’ rights and interests should be specified individually in the relevant feature or a separate explanation. The processing scope of the relevant capabilities is as follows, subject to the device you use and the features enabled:
Real-name or enterprise authentication: Feishu integrated login is provided with the cooperation of Party A;
Voice/image/screen capabilities: Used only with user authorization;
Location, payment, financial, or health-related capabilities: Not involved.
Please do not use DHwork to process Sensitive Personal Information unless you have confirmed the lawfulness, necessity, access controls, and recipients for the scenario and obtained the necessary authorization. Organizational Customers should also complete a Personal Information protection impact assessment or other applicable compliance procedures.
VI. Entrusted Processing, Sharing, Transfer, and Public Disclosure
6.1 Entrusted Processing
We may engage providers of cloud infrastructure, model services, identity authentication, message notifications, customer support, security monitoring, data storage, or professional services to assist in providing the Service. We will select entrusted parties with appropriate security capabilities, agree on processing purposes, duration, means, information types, protective responsibilities, deletion/return requirements, and liability for breach, and supervise them to ensure that they do not process information beyond the scope of the entrustment.
6.2 Sharing and Provision
We will not provide your Personal Information to unrelated third parties unless we obtain your consent, follow your instructions, it is necessary to perform a contract or fulfill statutory obligations, or applicable law otherwise permits it. Possible recipient categories and processing scopes are as follows:
| Recipient/Category | Information That May Be Received | Purpose of Use | Processing Role and Location | Explanation |
|---|---|---|---|---|
| Cloud infrastructure, hosting, and security service providers (such as Tencent Cloud) | Necessary portions of account information, service data, and logs | Hosting, storage, operations and maintenance, and protection | Usually entrusted parties; the location depends on the service deployment region | Processing only within the scope necessary to provide the corresponding service |
| Model service providers offered by the platform (such as Anthropic, OpenAI, Google, and ByteDance, depending on the selected model) | Input, context, task parameters, and necessary metadata needed to complete requests | Inference, content safety, or model capabilities | The specific role depends on the rules of the corresponding service; may involve overseas processing | Necessary information is transmitted according to the selected model and feature, subject to the corresponding service rules |
| PostHog usage analytics service | Email address, name (if available), application version, platform, and feature usage and feedback records | Product usage analysis and improvement | Analytics service provider; may involve overseas processing | The mobile version reports data after your consent; message bodies and file contents are not included |
| Third-party models, tools, data sources, or applications you actively connect | Requests, token/account authorization information, and corresponding content within the scope you authorize | Complete connection or automation tasks you initiate | Usually processed independently by the third party | Authorization page, least privilege, withdrawal entry point, and third-party policies |
| Organizational Customers and their Administrators | Member and audit information in organizational accounts, and Enterprise Data visible according to the configuration | Organizational management, security, collaboration, and compliance | The Organizational Customer is usually the processor | Scope of visibility and auditing, and export and retention policies |
| Professional advisers, auditors, regulators, or competent authorities | Information required by law or necessary for dispute handling | Fulfill statutory obligations, protect rights, or respond to lawful requests | Determined under applicable law | Request review, minimal disclosure, and record-keeping mechanisms |
When sharing Sensitive Personal Information, providing Personal Information overseas, or in other circumstances where separate consent is required by law, we will provide separate notice, obtain separate consent, conduct a Personal Information protection impact assessment, and complete other statutory procedures as required by law. If recipients or processing scopes change, we will update the explanations under Section IX of these Guidelines and obtain consent again where required by law.
6.3 Transfer, Merger, and Public Disclosure
If a merger, division, reorganization, acquisition, asset sale, bankruptcy liquidation, or similar event results in a transfer of Personal Information, we will inform you of the recipient’s name and contact details as required by law and require it to remain bound by these Guidelines or protective obligations no less protective than these Guidelines; if the purposes or means of processing change, we will obtain consent again or satisfy other applicable conditions as required by law.
We will not publicly disclose Personal Information unless we obtain separate consent, disclose it based on your active choice, or laws or regulations provide otherwise. Strict review and minimization measures should be applied to any public disclosure.
VII. Storage, Retention, and Security
7.1 Storage Locations and Retention Periods
We will retain Personal Information for the shortest period necessary to achieve the corresponding purposes and, after that period expires, the Service terminates, or you submit a valid deletion request, delete or anonymize it, or cease processing other than storage and necessary protection, in accordance with applicable law and technical capabilities. Any retention required by laws or regulations, taxation, auditing, dispute resolution, security incident investigations, or backup recovery will be limited to the necessary scope and duration.
The foregoing retention rules apply to account information, conversations/input/output, files and indexes, memory data, task and tool logs, security logs, support tickets, and backups. For Enterprise Data controlled by an organization, the agreement between the Organizational Customer and us should also be followed; you may use the contact details in Section X of these Guidelines to learn about the retention and processing arrangements for the corresponding data.
7.2 Security Measures
We will take reasonable administrative and technical measures based on risk, including access controls and least privilege, identity authentication, encryption in transit and at rest (where applicable), key and credential management, isolated environments, log auditing, vulnerability management, secure development, vendor management, employee confidentiality training, backups, and disaster recovery. No system can guarantee absolute security; please safeguard your passwords, tokens, and devices, and grant access permissions to agents cautiously.
If a security incident involving leakage, tampering, loss, or similar compromise of Personal Information occurs or may occur, we will activate emergency response mechanisms, take remedial measures, and report to or notify the competent authorities and affected individuals as required by law. Notifications may be made through in-service messages, email, SMS, telephone, announcements, or other reasonable means; where individual notification is difficult, we will use reasonable and effective public announcements.
VIII. Your Rights and How to Exercise Them
To the extent provided by applicable law, you may request access to, copies of, correction of, supplementation of, or deletion of Personal Information; withdraw consent; restrict or refuse processing; request explanations of processing rules; transfer Personal Information that meets the relevant conditions; close your account; and express views on automated decision-making or request human review (where applicable).
You may exercise the relevant rights through the following product entry points or the contact details in Section X of these Guidelines, subject to the features supported by the device you use:
Account information, passwords, and login methods: Unified access through your enterprise’s Feishu accounts is supported, or you may log in with an email account and password provisioned by an Administrator;
Conversations, input, output, files, and tasks: View, export, and delete;
Memory and personalization: Can be viewed, edited, turned off, and cleared in DHwork;
Authorized Connectors, third-party accounts, and tokens: DHwork authorization management and withdrawal entry points;
Device permissions: Manage through operating system permission settings and by setting authorized access paths in DHwork;
Account closure: You can delete your account in the App’s “Settings.” Enterprise organization accounts are deactivated upon submission of a request, and we will complete manual processing within 15 working days; deletion of default organization accounts takes effect immediately, and they can no longer log in;
Organizational Members: For Enterprise Data controlled by an Organizational Customer, please contact the Organizational Customer or Administrator first; we will assist within the scope of applicable law, contracts, and technical capabilities.
To protect your rights and those of others, we may verify your identity before responding to a request and may ask you to provide reasonably necessary information. We will process and respond within the time limit prescribed by law; if an extension is needed, we will explain the reasons. Where China’s Personal Information Protection Law applies, we will usually respond within 15 working days; in other regions, requests will be handled under applicable law. If you are dissatisfied with the outcome, you may complain to the relevant regulatory authority or seek remedies through the channels provided by applicable law.
IX. Updates to These Guidelines
We may revise these Guidelines because of changes in laws or regulations, adjustments to product features, changes in data processing activities, or updates to security measures. If a change will materially affect your Personal Information rights and interests, we will explain the effective date, major changes, and available options through reasonable means (such as prominent in-app notices, pop-ups, emails, or website announcements); where consent must be obtained again by law, we will obtain the corresponding consent before processing.
X. Contact Us
If you have questions, comments, complaints, or reports regarding these Guidelines, Personal Information processing, or the exercise of your rights, you may contact us through the following channels:
Contact email:
support@dhwork.netPersonal Information protection contact/email:
privacy@dhwork.net
Please describe your account identifier, the content of your request, and the necessary circumstances relevant to it, but do not send unnecessary Sensitive Personal Information by email. After completing identity verification, we will handle your request in accordance with applicable law and these Guidelines.