Zettlab Memo AI Feature Notice

Version date: June 3, 2026

This Notice applies to users outside Chinese Mainland who use Zettlab Memo AI chat, image/file understanding, NAS content processing, tool calls, Agent/Skill features, model selection, free credits, user-provided API keys and automation assistance.

Key Points

1. Zettlab Memo may support both local AI and cloud AI models. Data may leave your local device when you select a cloud model, enable remote access/sharing/connectors, send attachments to a model, publish Skills/templates or submit logs.

2. Content generated by AI-related services may be produced through generative artificial intelligence technologies by local models, cloud-based model services provided by us, third-party/cooperative model services, or user-supplied API keys/custom models, and does not represent the opinions or positions of Zettlab. Due to the inherent characteristics of artificial intelligence technologies, AI-generated outputs are probabilistic in nature and may contain errors, omissions, outdated information, fabricated content, or biases. The output content cannot be guaranteed to be identical or unique across different times or for different users. You are responsible for independently verifying such content and should not rely on it as the sole basis for decisions relating to medical, legal, financial, investment, personal safety, public safety, or other significant rights and interests. You shall bear the corresponding legal responsibilities for any judgments, decisions, or subsequent actions made based on such output content. Please use such services prudently, rationally, and in compliance with applicable laws and regulations.

3. Do not send identity documents, payment data, passwords, health data, precise location, unreleased business secrets, third-party private information or data you have no right to process to third-party AI models.

4. AI tool calls may read or write files, call local services, access networks, trigger automation or send requests to third-party services. Review tool permissions and data scope before enabling them.

1. AI Feature Scope

Depending on the product version, device capability, model configuration and service region, Zettlab Memo may provide:

1. text chat, summaries, Q&A, translation, rewriting, coding and creative assistance;

2. image, video, audio, document, code and attachment understanding;

3. NAS file Q&A, summaries, tagging suggestions, search assistance and content organization;

4. device status analysis, log explanation, troubleshooting and OTA assistance;

5. Agent personality, Memory, Skill, templates, scheduled tasks and automation suggestions;

6. connectors and tool calls such as Feishu, Slack, GitHub, Notion, Gmail, MCP services and custom tools;

7. Skill/template publishing review, security scanning and quality assessment;

8. model selection, credit usage, chat history, model configuration and user-provided API keys.

2. How AI Processing Works

When you use AI features, your prompts, context, chat history, attachments, file snippets, tool results, model parameters and request metadata may be sent to a local model or the cloud model you choose. We may perform security filtering, abuse detection, format conversion, compression, truncation and caching on input and output data, as well as record relevant logs. Where manual review is truly necessary, we will conduct such processing within the scope of compliance authorization, minimum necessity principle and access permission control.

Local models generally run on your device. Cloud models are processed by their providers and may be subject to their logging, retention, training, abuse monitoring, safety and security policies. If you configure your own API key or custom endpoint, you are responsible for selecting the provider and understanding its terms, privacy policy and legal implications.

3. Model Providers and Data Destinations

Actual models, providers, regions, retention and training settings depend on product configuration, service region and provider policies. Disabled models will not process your data.

Model/service	Region/scenario	Data processed	Training	Logs/retention	Cross-border	Status
Local model / local Agent	On-device chat, file Q&A, device assistance	prompts, outputs, attachments, file snippets, tool results	determined by local configuration	local chats/logs	usually no, unless remote access/sharing/log upload is enabled	depends on device
DeepSeek/Qwen/Volcengine	Cloud model, default model or review	prompts, outputs, attachments, context, parameters, metadata	per provider policy and enterprise settings	per provider/product configuration	may be processed in provider regions	configured/planned
Zhipu BigModel / GLM	Mainland-compatible model or review	prompts, outputs, attachments, context, metadata	per provider policy and enterprise settings	per provider/product configuration	may be processed in provider regions	configured/planned
OpenAI API / Anthropic / Google AI	Optional overseas models or user API keys	prompts, outputs, attachments, context, metadata	per provider policy and account settings	per provider policy	yes, may be outside your country/region	optional
User custom OpenAI-compatible endpoint	User configuration	data sent by the user	provider controlled	provider controlled	provider controlled	user responsibility
Third-party connectors / custom tools invoked by AI or Agent	Connector-enabled tool calls and automation	Connector account identifiers, scopes, tool inputs, messages/files/content snippets, tool results, runtime metadata; connector credentials for authentication only	Not an AI training use by Zettlab; selected provider/custom endpoint policy applies	Connector audit metadata where enabled; no intentional storage of connector secrets/tokens/full requests/full responses in audit records; runtime tokens temporary memory only	Depends on selected provider or endpoint	Optional; user-enabled

For links to the privacy policies of third-party AI model providers and information on their latest updates, please visit the Third-Party AI Models page.

4. Free Credits and Paid Features

Zettlab Memo may grant free AI credits to registered users. Credits may expire, may not stack, and may be changed, suspended, limited or withdrawn due to model cost, service strategy, abuse prevention, campaign rules or technical changes. Specific credit amount, validity, usage and refresh rules are shown in the product.

As of the date hereof, paid services are not currently available. If paid credits, membership, subscriptions or premium services are introduced later, pricing, billing cycle, cancellation and refund rules will be shown before purchase. 

5. Input, Output and High-Risk Uses

AI output may be wrong, incomplete, outdated, fabricated, biased, unsuitable for a purpose or not reproducible. You should independently verify output and should not use it as the sole basis for:

1. medical, legal, financial, investment, insurance or tax decisions;

2. personal safety, property safety, public safety or important rights decisions;

3. evaluation, ranking, screening, credit, hiring, education, insurance pricing or other significant automated decisions about individuals;

4. unlawful, infringing, harassing, discriminatory, fraudulent, malicious, cyberattack, privacy-invasive or regulatory-evasion activities.

6. Files, Images, NAS Content and Tool Calls

When you upload images/files or ask AI to process NAS content, the content may be read, compressed, converted, chunked, indexed, cached, sent to local services or sent to the selected AI provider. Images may be sent as files, URLs, base64 or data URLs depending on network conditions, model capability and implementation.

If AI features involve file read/write, device status reading, web access, command execution, third-party connectors, MCP services or automation, we will implement permission controls, logs and prompts where feasible. You are responsible for ensuring you have lawful rights to process the relevant files, devices, accounts and data.

Connector and tool-call data flow: if you enable a connector for an AI Agent/Skill, tool requests may include the minimum information needed to complete the action, such as selected account/connection identifiers, tool inputs, relevant messages/files/content snippets, tool results and runtime metadata. Connector credentials are used for authentication and are not intentionally included in AI prompts or connector audit logs. Third-party connector providers or custom endpoints may process the data under their own terms and privacy policies.

Before enabling or using a connector that shares personal data with a third-party service, the product should present the provider or endpoint, the categories of data involved and the purpose of the sharing, and obtain the authorization required by applicable law or platform rules. You can revoke or delete connectors where supported.

7. Safety, Labeling and Review

We may apply safety filtering, abuse monitoring, content blocking, account/function limits, automated review or human review to prompts, outputs, uploads, Skill/template packages and tool calls. If use violates law, platform rules or our terms, we may refuse generation, stop responses, limit features, suspend accounts, preserve evidence or report to authorities where required.

AI-generated synthetic content within Zettlab Memo will be labeled in the following manner: (i) when to label — labels shall be applied at the time of generation and remain attached to the content throughout its lifecycle within the platform; (ii) form of labeling — labels may take the form of visible watermarks, metadata tags, embedded provenance signals, or other technically appropriate indicators that identify the content as AI-generated; (iii) retention obligation upon external distribution — when you copy, download, export, share, or publicly publish such content, you are obligated to retain all applicable AI-generated content labels and must not remove, obscure, or alter them. Failure to retain such labels when distributing AI-generated content externally may constitute a violation of applicable laws, including transparency requirements under the EU AI Act and other relevant regulations.

8. Data Retention and Deletion

When you use AI features, we will collect and process AI sessions, inputs, outputs, attachments, model selections, tool call records, and usage quota records. The relevant data may be stored in your account, locally on your device, or in cloud services. Such data is collected and used for the following purposes: (i) Purposes necessary for service delivery: to display conversation history, enable continuous dialogue, provide security protection, support troubleshooting, and manage usage quotas; (ii) For AI feature improvement, subject to your separate authorization: your conversation data may be used to optimize AI features. You may grant or withdraw this authorization at any time, and withdrawal of authorization will not affect services already provided. Within the scope of product support, you may delete sessions, clear cache, unbind devices, or deactivate your account.

After account closure, we will stop retaining business data except where required for law, security audit, disputes or rights protection. Data processed by third-party AI providers is governed by their policies and account settings. We will disclose within the product the integrated providers, retention/training settings, and the scope of deletion assistance.

9. User Responsibilities

You shall bear full legal liability for all content uploaded, published or transmitted through this service and ensure that content you input, upload, authorize for processing or publish has a lawful source and legal basis. You must not use AI features to generate, distribute or facilitate illegal, infringing, fraudulent, harassing, discriminatory, malicious, cyberattack, regulatory-evasion, privacy-invasive or harmful content or conduct.

10. Contact

If you have any questions about AI features, model processing, generated content or deletion, contact support@zettlab.com. Upon receipt, we will promptly verify and process your request. Requests relating to personal information rights will be responded to within the time limits specified in the Zettlab Memo Privacy Policy (Non-Chinese Mainland Version).