Memmy Data Use and Privacy Notice

Last Updated: 2026-06-16

Effective Date: 2026-06-16

I. Overview

This Memmy Data Use and Privacy Notice (this "Notice") is intended to explain, in a clear and transparent manner, what information Memmy (the "Service") processes, why and how it is processed, with whom it is shared, and the rights you have in respect of your own information.

We adhere to the principles of "local-first, data minimization, and user control": in the current version, your core data remains on your device by default; we process information only to the extent necessary to provide and improve the Service; and we will use your content to improve our products or train our models only on the condition that you have proactively enabled the Improvement Program.

This Notice applies to the Memmy desktop client and its related services. It should be read together with the Memmy Terms of Service.

II. Where Your Data Is Stored

In the current version, we adopt a local-first architecture. By default:

your conversation history, memory data, configurations and similar information are stored on your local device;

under the Bring-Your-Own API Key mode, the API Key you configure is stored solely on your local device and will not be uploaded to our servers; and

memory data generated through cross-Agent collection is stored on your local device by default.

The data storage architecture may evolve with product iterations (for example, future support for cloud-based or hybrid storage modes). This Notice describes the current version; we will separately notify you of any relevant changes and update this Notice accordingly.

In the following scenarios, the relevant data will leave your device (see Sections IV and V for details): registration/login verification, the processing of requests to large language models, vector models and the like, and the use of data after you enable the Improvement Program.

III. Information We Collect

3.1 Account Registration Information (Registered Account Mode only)

Mainland China edition: mobile phone number (used for login verification, account identification, and necessary security risk control);

Overseas edition: email address (used for login verification, account identification, and necessary security risk control); and

the nickname you set (for display purposes only and may be left blank).

Purposes of collection: to create and maintain your account; to provide you with the various platform services offered to registered users (currently primarily platform large language model, vector model and similar services, and to meter their usage); to safeguard account and service security; and, where a complimentary-credit promotion applies, to grant you complimentary credits. Legal basis: necessity for the performance of the Memmy Terms of Service, and your consent.

The Bring-Your-Own API Key mode (Mode A) does not require registration, and we do not collect your mobile phone number or email address. Login and verification methods may be adjusted or added with version updates.

3.2 Basic Operational Statistics

In order to understand the basic operation and activity scale of the Service, the Service may collect the most basic operational statistics (such as software launch/exit counts) even if you have not enabled the Improvement Program.

3.3 Information Processed After Participating in "Help Us Improve" (the Improvement Program)

The Improvement Program is disabled by default. It corresponds to your "share data" selection in Settings:

If you have not enabled the Improvement Program: we will not use your conversation content, prompts, memory data or the like to improve our AI features or train our models.

If you proactively enable the Improvement Program: we may use and store your conversation content, prompts, Agent operation logs, memory data, feature-usage behavior, crash and error information, and other usage-related data, in order to improve our AI features and train our models.

Purposes of collection and use: to diagnose issues, improve features and stability, and enhance AI capabilities. Legal basis: your explicit consent. You may withdraw such consent at any time (see Section VIII); following withdrawal, we will cease any further collection and use of the foregoing data under the Improvement Program.

IV. Our Restraint Outside the Improvement Program

Regardless of whether you have enabled the Improvement Program, we exercise restraint with respect to the following information:

under the Bring-Your-Own API Key mode, your API Key is stored solely on your local device and will not be uploaded to our servers;

where you have not enabled the Improvement Program, we will not use your conversation content, memory data or the like to improve our products or train our models; and

we will not sell your personal information.

Note: whether conversation, memory and similar content is used for product improvement and model training depends on your choice regarding the Improvement Program; see Sections III and VIII for details.

V. Data Flows for Cross-Agent Memory Collection and Large Language Model / Vector Model Requests

5.1 Cross-Agent Memory Collection (Continuous/Incremental Scanning)

After you proactively grant collection authorization, the Service will read conversations of other Agents on your local device in order to distill memories:

memory distillation may be performed by a local model or a cloud-based model. In circumstances such as the use of Platform Tokens (Registered Account Mode), the conversation content being distilled will be transmitted to large language model, vector model and similar suppliers, or to our model proxy service, for processing;

the resulting memory data is stored on your local device by default, and you may disable collection or delete the data at any time in "Memory Management"; and

whether such content is further used to improve products or train models depends on your choice regarding the Improvement Program (see Section III).

5.2 Large Language Model / Vector Model Requests

Invoking large language models, vector models and the like is an inherent step in using AI services, and the request content you input will be processed by such suppliers:

Bring-Your-Own API Key mode: requests are sent by the Software directly to the supplier you have configured; we do not act as an intermediary proxy and do not store the request or response content; and

Registered Account Mode: when Platform Tokens are used, requests are routed through our model proxy service to large language model, vector model and similar suppliers. For purposes such as providing the Service and safeguarding stability and security, we may store the relevant request and response content; where you have not enabled the Improvement Program, we will not use such content to improve our products or train our models.

Please note that the processing of your request content by large language model, vector model and similar suppliers is governed by their respective privacy policies.

VI. Sharing and Entrusted Processing of Information

We do not sell your personal information. We share or entrust the processing of your information only in the following circumstances, and we require the relevant parties to fulfill corresponding data protection obligations:

Verification-code/communications service providers: to complete login verification (such as sending verification codes), we may provide the mobile phone number/email address to the relevant service provider;

Large language model, vector model and similar suppliers and the model proxy: to deliver AI capabilities, your request content is processed as described in Section 5.2;

Official-website analytics: our official website may use analytics/event-tracking tools to understand visits and downloads; and

Disclosure required by law: where required by laws and regulations, or where judicial or administrative authorities so require through statutory procedures, we may provide the relevant information in accordance with the law.

The above third parties are listed merely to illustrate the current categories of sharing/entrustment; specific service providers may change as the business is adjusted, and we will continue to require them to fulfill corresponding data protection obligations.

VII. Your Rights and Choices

You have the following rights and choices with respect to your information:

Disabling the Improvement Program / withdrawing consent: you may disable the Improvement Program at any time in "Settings"; after disabling, we will cease any further collection and use of the relevant data under the Improvement Program. You may also re-enable it at any time.

Managing memory collection: you may, at any time in "Memory Management," disable collection, disable injection, or delete the memory data already generated.

Deleting local data: you may at any time delete the data stored by the Software on your device in order to clear local conversations, memories and configurations.

Accessing/correcting account information: registered users may view and modify information such as their nickname in Settings.

Deregistration/logout: registered users may log out; to deregister an account, you may submit a request through the contact information in Section IX.

VIII. Minors

We attach importance to the protection of minors' personal information. If you are a minor, please use the Service with the consent and under the guidance of your guardian. If a guardian discovers that a minor has used the Service or provided personal information without consent, please contact us through Section IX, and we will handle the matter in accordance with the law.

IX. Policy Updates and Contact Information

9.1 We may update this Notice from time to time. Updates will be published through in-app notifications, announcements on the official website, or other reasonable means, and the "Last Updated" date at the top will be revised. We will provide prominent notice of material changes.

9.2 If you have any questions, comments or complaints regarding this Notice or the protection of personal information, please contact us:

Email: memmy@memtensor.cn