Login
- Released At: 13-02-2026
- Page Views:
- Downloads:
- Table of Contents
- Related Documents
-
H3C LinSeer Copilot User Guide
Copyright © 2026 New H3C Technologies Co., Ltd. All rights reserved.
No part of this manual may be reproduced or transmitted in any form or by any means without prior written consent of New H3C Technologies Co., Ltd.
Except for the trademarks of New H3C Technologies Co., Ltd., any trademarks that may be mentioned in this document are the property of their respective owners.
This document provides generic technical information, some of which might not be applicable to your products.
Contents
Features supported by LinSeer Copilot
LinSeer Copilot service configuration process
Log in to the LinSeer Copilot page
Guidance for interacting with front-end chat pages
Guidance for using back-end management pages
Resource permission management
Guidance for using the AI Copilot code plugin
Install the LinSeer Copilot code plugin
Configure the LinSeer Copilot code plugin
Overview
About this product
H3C LinSeer Copilot delivers a native AI experience by leveraging the LinSeer large language model and its LinSeer Hub capabilities. It simulates human intelligence to offer natural language processing, machine learning, and problem solving. This enables users to manage their daily tasks more intelligently, efficiently, and conveniently, and to resolve their questions. LinSeer Copilot permits integrating local knowledge bases to improve accuracy of retrieved information. Additionally, LinSeer Copilot offers diverse Q&A assistants and smart applications tailored to your specific service scenarios, helping you leverage the large language model capability to boost efficiency and reduce costs across various business processes.
Figure 1 Relationship between the large language model, LinSeer Hub, and LinSeer Copilot
Features supported by LinSeer Copilot
LinSeer Copilot supports multiple features, including but not limited to the following:
· Knowledge Q&A: Answers a wide range of questions, including daily life tips, historical events, work skills, and more.
· Language translation: Enables multilingual translation across different languages.
· Code programming: Understands, edits, and writes code.
· Mathematical reasoning: Possesses chain-of-thought reasoning capability to perform scientific and commonsense reasoning.
· Creative association: Generates creative content and makes associations in various styles and for multiple tasks, such as emails, official documents, and proposals.
· Auto creation: Generates PPT files, pictures, and Word documents automatically based on user needs.
· Search & summary: Uses an integrated search engine to retrieve user questions, analyzes the retrieved content, and generates a summary report.
Benefits
LinSeer Copilot delivers the following benefits:
· Unified entry for users to access AI services: LinSeer Copilot offers a convenient AI service access point. It supports both Web-based access and mobile endpoint-based access and can be seamlessly integrated into third-party applications. Users can perform complex configurations on LinSeer Hub and enjoy simple, intuitive model inference conversation services on LinSeer Copilot, enhancing user experience.
· Capability integration and scenario customization: LinSeer Copilot combines the model functions of LinSeer Hub with its extended features such as prompts, knowledge bases, and agents. LinSeer Copilot allows you to customize exclusive scenarios based on user requirements and presets various industry-specific assistants to meet diverse customer requirements across different industries.
LinSeer Copilot usage methods
In the current software version, users can access the LinSeer Copilot function through the following tools, as shown in Figure 2.
· LinSeer Copilot page: Users can access the LinSeer Copilot page through the Web interface or mobile endpoint to use all LinSeer Copilot functions and manage and configure LinSeer Copilot.
· Code plugin: LinSeer Copilot offers a plugin designed for code editors or integrated development environments. You can install and use this plugin in third-party software.
Figure 2 LinSeer Copilot usage methods
LinSeer Copilot page
Users primarily use LinSeer Copilot in the following ways:
· Interact with LinSeer Copilot using prompts. A prompt is input that a user gives to an AI assistant. It can be a text-based question, request, prompt, or other type of statement. The user's input prompt tells the AI assistant what task to perform or what question to answer, such as "Tell me today's weather" or "Help me draft a professional email reply." The AI assistant parses user needs based on these prompts and provides relevant information, executes tasks, or engages in further interaction with users.
Prompt design is critical to AI assistant performance. A well-designed prompt helps AI better understand user intent and deliver more relevant, higher-quality responses. Here are some tips for designing prompts:
¡ Define your goal: Before designing a prompt, determine what information you want from the AI assistant or what type of task you want it to perform.
¡ Define the role: Assign a specific role in the prompt to help the AI assistant better understand user intent and respond as expected for that role. For example, design a prompt like this: "As a travel consultant, recommend destinations and itineraries."
¡ Use clear language: Use direct and precise words. Avoid vague or ambiguous terms and unnecessary complexity.
¡ Structured prompt: Use structured prompts to help AI parse user intent for complex requests. For example, if you design a travel advisor, organize it by structure like "destination-date-activity type."
· Use LinSeer Copilot to automatically generate Word documents, PPT documents, or pictures, translate texts, or parse document descriptions for Q&A. These applications leverage large model capabilities to expand the AI assistant's applicable scenarios, delivering convenient and practical all-around services to users.
· Use AI search to find the required information on the Internet or in a knowledge base. AI search is an AI-powered search method. Unlike traditional search, it deeply understands user intent and delivers smarter search results for a better search experience.
To get a better search experience, use these tips when using AI search:
¡ Use clear language: Be direct and precise, and avoid vague or ambiguous words.
¡ Switch expressions: If the search results are unsatisfactory, try using different expressions to search again.
|
|
NOTE: When you access the LinSeer Copilot page from a mobile endpoint, you cannot use the apps and the AI search feature. You can only interact with LinSeer Copilot through prompts. |
LinSeer Copilot code plugin
After you install the LinSeer Copilot code plugin in a third-party code editor or integrated development environment, you can directly use the plugin to intelligently generate code, understand and comment code, or engage in intelligent Q&A through conversations with LinSeer Copilot.
LinSeer Copilot service configuration process
As shown in Figure 3, to use the LinSeer Copilot service on the front-end pages, complete the following service configuration on the back-end management pages of LinSeer Copilot:
1. Obtain a license: Access the System > License > License Info page, and configure the license server information to obtain LinSeer Copilot licenses from the license server. If you do not do that, you cannot use LinSeer Copilot.
2. Synchronize resources: LinSeer Copilot does not include resources like large language models or knowledge bases by default. Follow these steps to synchronize large language models, knowledge bases, plugins, and agents from LinSeer Hub.
a. Access the Design > Permission page, and configure the LinSeer Copilot parameters, including the LinSeer Hub URI for LinSeer Copilot to request resources from LinSeer Hub.
|
|
TIP: When you are unsure about a parameter's function or its impact on the system, keep the default settings, and avoid changing the name and key value of that parameter to prevent errors. |
b. Access the System > Scheduled Tasks page. Configure scheduled tasks to automatically synchronize resources from LinSeer Hub. Alternatively, manually execute resource synchronization tasks once.
After resources are synchronized, you can view and use the large language models, knowledge bases, plugins, and agents synchronized from LinSeer Hub on LinSeer Copilot.
3. Manage permissions: Manage page operation permissions and resource permissions.
¡ Page operation permissions: Specify the LinSeer Copilot pages that users can access and operate. To set page operation permissions, access the System > Users/ Roles/Departments page, and then configure users, roles, and departments.
¡ Resource permissions: Specify the resources that users can use. To set resource permissions, follow these steps:
- Access the System > Users/Roles page, create users and roles, and assign roles to users.
- Access the Resources > Resource Permissions page, and configure resource access permissions for each role.
4. (Optional.) Manage resources: LinSeer Copilot resources include models, knowledge bases, plugins, and assistants. After you configure and manage resources on the back-end management pages of LinSeer Copilot, you can view and use these resources on the front-end pages of LinSeer Copilot. You can manage different resources as follows.
¡ Prompts: A prompt is a template for user questions when users interact with LinSeer Copilot. By using prompts, users can ask more precise questions in specific fields, enhancing the effectiveness of LinSeer Copilot. To create and manage prompts, access the Resources > Prompts page.
¡ Assistants: An assistant provides intelligent Q&A services for specific scenarios or fields. An assistant presets knowledge bases and prompts for specific scenarios or fields to deliver better AI-powered Q&A services. To create an assistant, access the Resources > Assistants page and manually create an assistant. Alternatively, access the Resources > Agents page and release an agent synchronized from LinSeer Hub as an assistant. Access the Resources > Digital Human page. Click Add. In the window that opens, set the digital human image and voice, and then bind the digital human to an assistant. After you bind the assistant to the digital human, the digital human's image appears on the chat page and the assistant responds with the digital human's voice function when you use the assistant.
¡ Models: Access the Resources > Models page, view the models synchronized from LinSeer Hub, and edit model parameters.
¡ Knowledge bases: Access the Resources > Knowledge Base page, and view the knowledge bases synchronized from LinSeer Hub.
¡ Plugins: Access the Resources > Plugins page. On this page, you can view plugins synchronized from LinSeer Hub and plugins customized on LinSeer Copilot, and specify whether a plugin can be used.
¡ Agent: An agent is an AI system built on large-scale pre-trained models, simulating human behaviors to perform tasks such as natural language understanding, decision support, and interaction. An agent can be applied in various scenarios, such as customer service, education, and consulting, providing users with services like Q&A, information retrieval, and data analysis. On the Resource Management > Agent Management page, you can synchronize agent data from the LinSeer Hub platform and publish the synchronized agents as AI assistants that support prompt interaction.
¡ Digit human: A digital human is a virtual character in the AI assistant, designed to enhance user interaction with vividness and a personalized experience. Once the AI assistant is bound to a digital human, it can display the digital human's image in the conversion interface. In addition, the AI assistant can answer user questions not only through text but also using the digital human's voice feature for spoken responses. You can customize digit human on the Resource Management > Digit Human page.
¡ Apps: Access the Resources > Apps page. On this page, you can manage LinSeer Copilot's smart apps, including smart writing, smart reading, translation, PPT, and draw.
¡ AI search: Access the Resources > AI Search page. On this page, you can set parameters for the AI search feature, including search mode, model settings, and recommended questions.
5. (Optional.) Release assistants: Access the System > Apps page, and release an assistant as a standalone app. Users can access and use the assistant through its released URL address.
6. (Optional.) Back up data: As a best practice, regularly back up data on the System > Data Backup page to prevent data loss when you are configuring LinSeer Copilot.
Figure 3 Recommended service configuration process
Log in to the LinSeer Copilot page
After you deploy LinSeer Copilot, enter its login address in the address bar of the browser to access the login page, as shown in Figure 4. The login address format is https://ip_address:30088 or http://ip_address:30099. The ip_address parameter represents the northbound service VIP of the LinSeer Copilot cluster as the login IP. The default username and password are admin and Admin@h3c, respectively.
Figure 4 LinSeer Copilot login page
LinSeer Copilot provides the front-end chat pages and the back-end management pages.
· Front-end pages: Users chat, perform AI searches, and use apps on these pages.
· Back-end management pages: On these pages, the administrators can manage resources available for front-end LinSeer Copilot, such as prompts, assistants, knowledge bases, plugins, and agents. The administrators can also configure the system appearance and parameters for both the back end and LinSeer Copilot.
After you enter your username and password,
click Login Now to access the front-end chat pages of
LinSeer Copilot, as shown in Figure 5. To change
the LinSeer Copilot display style, click the palette icon 
in the upper-right corner. To perform the
following tasks, click the user icon in the upper-right corner:
· View system information.
· Provide feedback.
· Access the back-end management pages, as shown in Figure 6.
· View the current user information.
· Log out.
Figure 5 Front-end chat pages of LinSeer Copilot
Figure 6 Back-end management pages of LinSeer Copilot
Guidance for interacting with front-end chat pages
Chat
About this feature
The Chat page uses large models for intelligent Q&A. After logging in, users can directly access the Chat page.
Tasks
· Customize these AI conversation settings to better match your specific needs:
¡ Large model: Select an inference service deployed by using the large model capability. Switch between different large models in the top-left corner of the page to achieve varied generation effects and capabilities.
¡ Knowledge base: Select the knowledge base to mount. A knowledge base aggregates structured knowledge in a specific domain or topic. After mounting, it enables real-time queries, intelligent retrieval, and reasoning based on the knowledge, enhancing the AI assistant's capability for professional Q&A and content generation.
¡ Online Search: Click the 
icon in the dialog
box to enable real-time online retrieval of the latest Internet information. The
search results are then combined with the large model's reasoning capabilities to
enhance both the timeliness and accuracy of responses.
¡ File Upload: Click the 
icon in the dialog
box to upload local documents. The large model can then provide context-aware and
targeted responses based on the file content.
¡ Intelligent Applications: Currently supports intelligent PPT applications. Click the <PPT> icon, and the AI assistant will automatically generate structured PPT content based on user Prompts for efficient content creation.
¡ Set prompt as recommended: Click the 
icon in
the dialog box to select the system's preset recommended prompt template,
enabling one-click efficient questioning and improving conversation efficiency.
¡ Conversion mode: If you select Multi-Round Conversation, the inference service will use the latest few question-answer pairs as context to respond. Without this option, the inference service will answer the user's question independently, without relying on past conversations.
If the above features are not available, configure the related parameters on the back end management > resource management > Webpage configuration page.
· Create a new conversation
Click New Conversation or switch the large model to create a new conversation. The AI assistant extracts content from the first question as the conversation title.
In a new conversation, you cannot create another conversation.
· Manage recommended prompts
Enter a forward slash (/) in the question
input box, then click View More; or click the 
icon in the dialog box. On Recommended Instructions page, you can favorite,
copy, and use prompts.
¡ Click Copy to copy a prompt to the clipboard. You can then paste it into the conversation or other locations.
¡ Click Favorite to save the prompt. Use a forward slash (/) in the question input box to access your saved prompts.
¡ Click Use to return to the previous conversation window. The prompt will be copied to the question input box.
· Invoke favorite recommended prompts
Enter a forward slash (/) in the question input box to access your favorite prompts. This allows users to quickly select prompts to boost efficiency, reduce input errors, and improve prompt accuracy.
· Copy conversation content or regenerate responses
After the large model completes its
response, you can click the 
icon to copy the conversation content or click the 
icon
to have the large model regenerate the answer.
· Provide feedback on responses
After the large model completes its
response, you can click the 
icon to give positive feedback or click the 
icon to
give negative feedback. The feedback results will be sent to the southbound LinSeer
enabling platform for analysis.
· View knowledge base retrieval results
If the knowledge base is selected during the conversation, the system will display the original text retrieved from the knowledge base based on the question on the right side of the page. You can view or download it.
· Export conversations
Click the floating Export Conversation icon on the right, and select the questions and answers to export. The system will save them to an Excel file.
· Chat History
The left panel of the chat interface displays your past conversations, including both regular chats and AI-assisted chats. You can:
¡ Resume any conversation by clicking on it to view the full chat history and continue the dialogue.
¡ Delete a specific conversation using the delete icon.
AI search
About this feature
Perform this task to use AI technologies to search the Internet for the required content.
Figure 7 AI search
Tasks
1. Click AI Search in the left menu bar to access the AI search page.
2. Enter the question you want to search in the input box, and then click the Send icon. The system will search for answers based on the input question.
Below the input box, you will find recommended questions. Click any question to search with its suggested text.
Click Online search in the dialog box to select whether to use the system-configured search engine for online searches. The search engine can be configured on the back end management > AI search management page.
Click Thinking in the dialog box to select whether to enable deep retrieval. Regular retrieval provides faster responses but might lack details, boundary conditions, and counterexamples. Deep retrieval tends to provide structured answers with bullet points or step-by-step explanations, offering more detailed reasoning paths and analysis of pros and cons.
3. During AI search, click Stop Generating to stop searching for answers. After the search finishes, the system analyzes the retrieved content and generates a summarized and organized answer.
4. After AI search generates an answer, you can perform the following tasks:
¡ View the reference sources for the retrieval results on the right side of the page.
¡ Click the outline tab on the right to view a list of questions for further detailed searches based on the original query. You can click items in the outline to conduct additional AI searches.
¡ To ask further questions about the answer generated by AI search, click Follow Up, enter your question in the dialog box that opens, and then click Send.
5. View and follow up on the search history and past Q&A in the search records section at the left.
Applications
About this feature
An AI assistant not only provides text-based AI Q&A but also delivers a range of smart office applications using AI models. This expands its applicable scenarios and offers users a more convenient, practical, and integrated AI tool.
Currently, an AI assistant supports the following applications:
· Intelligent PPT: Automatically generate a presentation based on the subject and description you provide.
· Intelligent document: Automatically generate a Word document based on user-specified purposes or refine and enhance the text content provided by users.
· Intelligent translation: Translate content into multiple languages.
· Document reading: Use large models to parse the content of uploaded documents. After parsing, users can ask questions about the document content.
· LinSeer multimodal service: Use large models to intelligently generate pictures.
Figure 8 Applications supported by LinSeer Copilot
Tasks
· Intelligent PPT
In the left menu bar, select PPT to enter the PPT document creation page. The AI assistant creates a PPT document as follows:
a. Enter the PPT subject.
b. (Optional.) Upload reference documents to generate PPT content or enter reference text content for generating PPT content.
c. (Optional.) Enter the speaker role and the PPT report type.
d. Select the PPT theme template. By default, the dark theme template is used.
e. Click Generate Outline to create a PPT outline. During the outline generation process, you can stop it.
f. In the generated PPT outline, click the title to edit it, or click the Add or Delete icon to add or delete the outline content. If you are unsatisfied with the outline result, click Regenerate to create a new PPT outline.
g. Click Generate Preview PPT,
and the system generates a preview PPT based on the outline. If unsatisfied with
the PPT, you can modify the outline or click Regenerate
to generate a new outline. Modified outlines can be applied by clicking Apply Outline to update the preview PPT. Click the 
icon to
download the PPT locally.
h. Click the 
icon to save the
generated PPT in the system. You can view the generated document on the My Works tab. You can click Download
to download a document to the local device or click Delete
to delete a document from the system. Click Preview
to preview the PPT.
Figure 9 LinSeer Copilot intelligent PPT
· Intelligent document
In the left menu bar, select Smart Writing to enter the Word document creation page. Word document creation currently supports content generation and content processing.
¡ Content generation
Click the Content Generation tab on the right, select the document type and purpose, enter the subject details, and then click Generate Body to create Word content.
During the generation, you can click Stop to stop the process.
After generating the Word document, if you are unsatisfied with the result, click Regenerate to create a new document with the same subject but reorganized content. To generate other documents, click Back to return to the content generation tab.
Exceptionally, for the Outline function, click Generate Outline to create an outline. Modify the outline by clicking titles, add, or delete icons. If unsatisfied, click Change for a new outline. If satisfied, click Generate Body to create the full document.
¡ Content processing
In a Word document with existing content, if you select any text paragraph, the function panel opens. You can use the following features to edit the text:
- Polish: Improve text details to enhance readability, flow, and expression.
- Expand: Add detailed descriptions and extensions to the original text to enhance its richness and detail.
- Simplify: Reduce and streamline the text to make it clearer and more concise.
- Continue: Add new content to the original text to extend its ideas and subject.
After processing the content, the system shows the processed text under the corresponding function bar on the Content Processing tab on the right. To add the processed text at the cursor position, click Insert. To overwrite the selected text with the processed content, click Replace.
During the generation, you can click Stop to stop the process.
The left panel lists all generated Word documents. You can click Download to save a document to the local device or click Delete to delete a document from the system.
To save the current document in the system, click Save in the upper-right corner. To download the current document to the local device, click Export.
· Intelligent Translation
In the left menu bar, select Translation to enter the content translation page. This application currently supports text translation and document translation.
¡ Text translation
Click the Text Translation tab, paste the text to translate into the left text box, select the target language on the right, and then click Translate.
¡ Document translation
Click the Document Translation tab, click Select Document, and upload the document to translate from your local device. Select the target language above, and then click Start. The system then starts to translate the document. The translation result appears in the History Records tab. You can download the translated document to the local device or delete it from the system.
In document translation, you can also specify terms for reference. For example, if you specify the source term as Personal Computer and the target term as PC, the system will translate all the Personal Computer in the document to PC. The terms must be organized in a glossary. During translation, the system uses the selected glossary for reference. You can edit and delete terms in the glossary, import external glossaries, or export the current glossary to the local device.
Figure 10 LinSeer Copilot intelligent translation
· Document reading
In the left menu bar, select Smart Reading to access the document-based Q&A page. You can use the Q&A feature based on document content as follows:
i. To start a conversation, click New Q & A, and then click Upload document to Start Q & A to upload the documents to read to the system.
j. After uploading the documents, click Start Q & A to begin question-and-answer interactions based on the document content.
k. The Document Q & A tab opens. Wait for the system to finish document parsing.
l. Enter your question in the right-side box to start a conversation.
m. Select a document from the My Chats tab on the left to start a conversation about the history document.
n. On the History Documents tab, you can view and manage documents uploaded to the system. On the History Documents tab, to start a conversation about a history document, click Read in the Actions column for the document. To delete a history document, click Delete in the Actions column for the document on the History Documents tab. To delete multiple history documents in bulk, select the documents, and then click Bulk Delete at the top of the section.
o. On New Q & A page, to start a conversation about multiple history documents, click Select History Document, select the target documents in the dialog box that opens, and then click OK.
Figure 11 LinSeer Copilot document reading
· LinSeer Multimodal Service
On the application selection page, click LinSeer Multimodal Service to enter the picture generation page. You can use the LinSeer multimodal service as follows:
p. Enter the text description of the picture you want to generate.
q. (Optional.) Enter unwanted content to exclude it from the pictures to generate.
r. Select a large model service.
s. (Optional.) Select the number of pictures to generate based on the text description.
t. (Optional.) Adjust the ratio and pixels of the pictures to generate.
Assistant center
About this feature
The page displays ready-to-use assistant templates. These templates provide complete prompt templates, knowledge base information, or workflow information to help users get more accurate answers or execute specific tasks. For example, the Emergency Plan Assistant predefines the role as an emergency rescue expert. The task is to draft a comprehensive plan for incident response, rescue, and resource allocation based on provided details and materials. The requirement is to write it according to professional emergency protocols and standards. With the Emergency Plan Assistant, users simply enter an event (such as blizzard weather or school commutes for K-12 students) in the question input box. The assistant template automatically converts it into an effective prompt to generate an emergency response plan for the event.
Figure 12 Assistant center
Tasks
Select an assistant template: After selecting the desired assistant template based on actual needs, click it to start a conversation based on that template.
User center
About this feature
After logging in, you can view and edit your profile in the user center.
Figure 13 User center
Tasks
Click the user icon in the bottom-left corner, and select User Center from the dropdown menu. On this page, you can view and update your basic information and change your password.
Feedback
About this feature
If you encounter any issues while using an AI assistant or have suggestions for improvement, please feel free to share your feedback through our Feedback feature. We highly value your feedback and will use your valuable input to continuously improve our service.
Tasks
Click the user icon in the bottom-left corner, and select Feedback. On this page, enter the feedback title, comment type, description, and issue screenshot, and then click OK to submit your feedback.
Figure 14 Feedback
Guidance for using back-end management pages
Resource management
Prompts
About this feature
Prompts are equivalent to question templates that users can follow when using the AI assistant. With prompts, users can ask more precise questions in specific areas, enhancing the effectiveness of the AI assistant.
The Prompts page manages prompts of the AI assistant. On this page, you can categorize, add, edit, delete, release, and remove prompts, as well as configure or cancel trending settings.
Figure 15 Prompts
Tasks
· Add a prompt category
Click Add Category. In the dialog box that opens, enter the category name and click OK. The newly added category will be displayed in the list on the left of the page.
Figure 16 Add a prompt category
· Edit a prompt category
To edit a prompt category, click the edit icon for that prompt category in the category list on the left of the page. You can edit the category name.
· Delete a prompt category
To delete a prompt category, click the delete icon for that prompt category in the category list on the left of the page.
· Add a prompt
a. Click Add.
b. In the dialog box that opens, specify the prompt name, select a prompt category, and enter the prompt content (AI assistant question template).
c. For users to use this prompt, select Listed from the Status field.
d. Click OK.
· Edit a prompt
To edit a prompt, click Edit in the Actions column for that prompt.
· Delete prompts
Use one of the following modes to delete prompts:
¡ To delete a prompt, click Delete in the Actions column for that prompt.
¡ To delete prompts in bulk, select the check boxes for the target prompts, and then click Delete on the top of the prompt list.
· Release or remove prompts
Select the check boxes for the target prompts, and then click Release or Removal on the top of the prompt list to release or remove these prompts, respectively.
· Set trending prompts
Select the check boxes for the target prompts, and then click Set as Trending on the top of the prompt list. When you create a new AI assistant conversation window, trending prompts will display as recommended prompts on the Web interface. You can select prompts from these trending prompts for quick Q&A.
To cancel a trending prompt, click Cancel Trending Setting in the Actions column for that trending prompt.
· Filter prompts
In the search area on the top of the page, enter and select search criteria and click Search. The page will then display only the prompts that match the criteria. To clear all search criteria and display all prompts, click Reset.
Assistants
About this feature
An assistant helps you ask better questions in specific scenarios or fields. After assistants are deployed, you can use the specified assistant for conversations on the LinSeer Copilot page. Then, LinSeer Copilot will respond in the scenarios or fields specified by the assistant. Additionally, you can preset questions in an assistant. Then, you can directly ask these preset questions to quickly get answers.
On the assistants page, you can categorize, create, edit, and delete assistants, and set recommended assistants.
Figure 17 Assistants
Restrictions and guidelines
· You can use an assistant in the Q&A feature of LinSeer Copilot only when the assistant is in listed state and authorized for your use on the Resource Permissions page.
· Assistants bound to apps cannot be deleted.
Tasks
· Add an assistant category
Click Add. In the dialog box that opens, enter a category name, and then click OK. The assistant categories you create are displayed as a list on the left of the page.
· Edit an assistant category
In the assistant category list on the left of the page, click the Edit icon for an assistant category to change its category name.
· Delete an assistant category
In the assistant category list on the left of the page, click the Delete icon for an assistant category to delete that assistant category.
· Create an assistant based on a URL
Click Create Assistant above the assistant list, select Create Assistant Based on URL, and then click OK to open the assistant creation page.
Create tab configuration:
a. Enter the assistant name and description, and select the assistant category.
b. In the URL address text box, enter the URL address. When users use this assistant, the system will directly access this URL.
c. In the state dropdown list, select the assistant listing state.
Configuration tab settings:
d. In the device configuration section, select whether the assistant is intended for PC or mobile use.
e. Use configuration items such as icons, avatars, banners, interface themes, and message backgrounds to customize the assistant appearance.
On the preview tab on the right, you can preview the assistant appearance as configured.
· Create an assistant
Click Create Assistant above the assistant list. Select Create Assistant Based on Knowledge Base, and then click OK to open the assistant creation page.
Create Tab Configuration:
f. In the dialog box that opens, enter an assistant name, provide a description, and select a category.
g. In the Prompt field, enter the prefix prompt for the assistant. The prefix prompt defines the context within which the large language model generates responses. For example, if you set the prompt to during safe production, you don’t need to specify the context when asking the assistant questions. The inference service will automatically use during safe production as the scenario to answer.
h. (Optional.) Select a knowledge base from the Knowledge Base list. The inference service will then retrieve answers from the selected knowledge base to provide more accurate answers to your questions.
i. (Optional.) Set TOP-K and searchScore
- TOP-K: During each knowledge base retrieval, returns the top K most relevant documents or snippets to the user's question.
- searchScore: The similarity or relevance threshold for retrieval results. Only documents with a relevance score equal to or higher than searchScore will be returned; content below this threshold is filtered out. Higher values yield more relevant results.
j. From the Model Selection list, select the large language model inference service for the assistant to use.
k. In the MCP service dropdown, select the MCP service for the assistant.
l. In the Recommended Questions field, enter preset questions for the assistant. Then, you can directly ask these questions. You can create up to two recommended questions.
m. From the Status list, select the releasing status of the assistant.
Configuration Tab Settings:
n. Under the Device configuration, choose whether the assistant is for PC or mobile terminal.
o. Customize the assistant's appearance using configuration items such as icon, profile picture, banner, theme, assistant message background, and user message background.
Preview the assistant's appearance in the Show Effect tab on the right and test its functionality in real-time via Q&A.
Figure 18 Create an assistant
· Edit an assistant
To edit the configuration of an assistant, click Edit in the Actions column for that assistant on the assistant list.
· Delete an assistant
To delete an assistant, click Delete in the Actions column for that assistant on the assistant list.
· Release or remove assistants
Select one or more assistants in the first column of the assistant list, and then click Release or Removal above the list to bulk release or remove the selected assistants.
· Search for assistants
In the search bar at the top of the page, enter and select search criteria, and then click Search. Then, the page will display only the assistants that match the search criteria. To clear all search criteria and display all assistants, click Reset.
Resource permission management
About this feature
Use this page to manage user role permissions for logging in to LinSeer Copilot, including viewing and editing permissions.
Figure 19 Resource permission management
Tasks
· View permissions
Click View Permissions in the Actions column for a role to view the role's assistant, model, knowledge base, and application permissions.
· Edit permissions
Click Edit in the Actions column for a role. In the dialog box that opens, edit the role's assistant, model, knowledge base, and application permissions as needed.
· Query roles
Enter your search criteria in the search bars in the top of the page, and then click Search. The page will only display roles that match your search criteria. To clear the search criteria and display all roles, click Reset.
Models
About this feature
The Models page is responsible for synchronizing model data from LinSeer Hub and managing the synchronized models. This page supports functions such as model synchronization, model retrieval by name, and editing model information.
Figure 20 Models
Tasks
· Synchronize models
The AI assistant system synchronizes model data from LinSeer Hub every five minutes by default. On the Scheduled Tasks page under System, you can modify the synchronization cycle. You can also immediately execute a model synchronization task in the following ways:
¡ On the Models page, click Model Synchronization.
¡ On the Scheduled Tasks page under System, click Execute Once in the actions column for the model synchronization task.
· Filter models
To obtain information about a model, enter the name of that model in the search area and click Search.
To clear all search criteria, click Reset.
· Edit a model
To edit a model:
a. Click Edit in the Actions for that model.
b. Set the following parameters:
- Temperature: Sets a temperature to control the randomness or determinism of the generation process. A higher temperature value will make the output more random and diverse, while a lower temperature value will make the output more focused and deterministic.
- Top_P: Set a value for this field. The smaller the Top_P value, the more certain the result. The greater the Top_P value, the more random the result.
The model name, inference name, and model version cannot be edited.
c. Click OK.
Knowledge base
About this feature
The Knowledge Base page is responsible for synchronizing knowledge bases from LinSeer Hub and displaying the synchronized knowledge base information. This page supports functions such as knowledge base synchronization, retrieval by name, and viewing knowledge base details.
Figure 21 Knowledge base
Tasks
· Knowledge base synchronization
The AI assistant system synchronizes knowledge base data from LinSeer Hub every five minutes by default. On the Scheduled Tasks page under System, you can modify the synchronization cycle. You can also immediately execute a knowledge base synchronization task in the following ways:
¡ On the Knowledge Base page, click Knowledge Base Sync.
¡ On the Scheduled Tasks page under System, click Execute Once in the actions column for the knowledge base synchronization task.
· Search for knowledge bases
Enter the knowledge base name in the search bar and click Search.
To clear the search criteria, click Reset.
· View knowledge base details
Click Details in the Actions column for the target knowledge base to view its name, description, response mode, and prompt.
Agents
About this feature
An agent is an AI system built on large-scale pre-trained models, simulating human behaviors to perform tasks such as natural language understanding, decision support, and interaction. An agent can be applied in various scenarios, such as customer service, education, and consulting, providing users with services like Q&A, information retrieval, and data analysis.
The Agents page is responsible for synchronizing agent data from LinSeer Hub and releasing the synchronized agents as assistants that support prompt interactions, offering services to users.
Figure 22 Agents
Tasks
· Agent synchronization
The AI assistant system synchronizes agent data from LinSeer Hub every five minutes by default. On the Scheduled Tasks page under System, you can modify the synchronization cycle. You can also immediately execute a agent synchronization task in the following ways:
¡ On the Agents page, click Agent Synchronization.
¡ On the Scheduled Tasks page under System, click Execute Once in the actions column for the agent synchronization task.
· View agent information
The Agents page displays information about the agents synchronized from LinSeer Hub to the AI assistant system in a list, including an agent's name, description, and guiding words.
· Released an agent as an assistant
To release an agent as an assistant, click Released as Assistant. In the dialog box that opens, perform the following operations:
a. Enter the assistant name and description.
b. Select the assistant type. You can only select an existing assistant type.
c. (Optional.) Enter recommended questions. Users can directly use these recommended questions when an agent is deployed as an assistant.
· View assistant information corresponding to an agent
Click View Assistant in the Actions column for an agent. You can view information about the assistant corresponding to that agent. This includes the assistant name, feature description (guiding words of the agent), assistant type, and recommended questions.
Design
About this feature
This feature helps you manage and personalize LinSeer Copilot's appearance and configure some conversation parameters.
Figure 23 Design
Tasks
Click Edit to edit system settings, conversation settings, and model conversation length limit.
· Configure system settings
In the System Settings area:
¡ System Name refers to the name shown in the upper left corner of LinSeer Copilot's homepage.
¡ Welcome Message refers to the guiding words displayed in the conversation window before the conversation starts.
¡ Copyright refers to the information displayed on the About page after clicking the user icon in the upper right corner of the homepage.
¡ System Logo refers to the icon in the upper left corner of LinSeer Copilot's homepage.
¡ Profile Picture refers to the image that the large model uses when it responds to questions in conversations.
¡ Login Page Background refers to the picture displayed on the login page when users access LinSeer Copilot.
You can edit the settings as needed.
· Configure conversation settings
In the Conversation Settings area, you can choose whether to show the model, knowledge base, conversation mode, and search engine options on LinSeer Copilot's chat page. You can also set the maximum number of conversation turns in this area.
· Configure the model conversation length limit
In the Model Conversation Length Limit area, the administrator can set the maximum number of characters users can input per query for each available model.
Digital human management
About this feature
A digital human is a virtual character in the AI assistant, designed to enhance user interaction with vividness and a personalized experience. Once the AI assistant is bound to a digital human, it can display the digital human's image in the conversion interface. In addition, the AI assistant can answer user questions not only through text but also using the digital human's voice feature for spoken responses. By mimicking real human appearance and voice, the digital human enables more natural conversations with users, offering a more immersive service experience.
On the digital human management page, users can select a digital human's appearance and voice audio, set the background image for the conversation interface, and specify the AI assistant to be bound to a digital human.
Figure 24 Digital human management
Tasks
· Add a digital human
Click Add. Then, set the digital human's name, connection mode, connection URL, appearance, and character voice, and then click OK.
· Edit a digital human
Click Edit in the Actions column for a digital human, and then edit the digital human's name, connection mode, connection URL, image, and character voice.
· Delete a digital human
To delete a digital human, click Delete in the Actions column for that digital human.
Parameters
Table 1 Parameters for adding a digital human
|
Parameter |
Description |
|
Digital Human Name |
Name of the digital human. |
|
Connection Mode |
Deployment and access mode for a digital human, including: · Public Cloud: The digital human service is deployed on cloud platforms accessible via the Internet. These cloud platforms are operated and managed by third-party cloud service providers. · Private Cloud: The digital human service is deployed in a restricted, private cloud environment, suitable for scenarios with higher data privacy and security requirements. |
|
appKey |
Unique identifier assigned by the digital human service provider to identify the application accessing the digital human service. The AI assistant includes an appKey in the requests it sends, and the digital human service uses this appKey to authenticate the AI assistant. This parameter is required only when you specify the public cloud connection mode. |
|
appSecret |
Secret key paired with the appKey to enhance the security of API requests. The appSecret is a string of characters that must be kept confidential to ensure the integrity of requests and prevent tampering. This parameter is required only when you specify the public cloud connection mode. |
|
Account |
Account that the AI assistant uses to access the digital human service, used for logging in to the digital human management system or service platform. This parameter is required only when you specify the private cloud connection mode. |
|
Password |
Password that the AI assistant uses to access the digital human service, used for verifying the AI assistant. This parameter is required only when you specify the private cloud connection mode. |
|
Digital Human Image |
Select a virtual avatar that meets your requirements from the digital human library provided by the public cloud or private cloud. These virtual avatars might feature diverse characteristics, such as gender, age, and clothing styles. |
|
Character Voice |
Select a character voice for the digital human. |
|
Bind Assistant |
Specify the AI assistant to be bound to the digital human. |
|
Digital Human Background |
Background image of the digital human conversation interface. |
Apps
About this feature
This page manages intelligent applications of the AI assistant, including determining whether applications are released, and editing the inference service used by an application.
Figure 25 Apps
Restrictions and guidelines
Only the applications that are released can be used by users.
LinSeer draw will automatically activate after you install the license. Other applications do not require activation.
Tasks
Click Edit in the Actions column for the target application, and perform the following tasks:
· Edit the application description, listing state, and the model used by the application.
· Set the temperature (sampling temperature) and Top-P (nucleus sampling) parameters for the application.
¡ Temperature: Controls the randomness or determinism of the generation process. A higher Temperature value makes the output more random and diverse, while a lower Temperature value makes the output more focused and deterministic.
¡ Top-P: Affects the determinism of the results. A smaller Top-P value makes the results more deterministic, while a larger value makes the results more random.
AI search
About this feature
On this page, you can set parameters for the AI search feature, including the search mode, model settings, and recommended questions.
Figure 26 AI search
Restrictions and guidelines
When you select the hybrid search mode, the AI search feature retrieves answers only from knowledge bases if the knowledge base threshold is higher than the online search threshold. If the knowledge base threshold you set does not exceed the online search threshold, the AI search feature retrieves answers from both the selected knowledge bases and the Internet, and then summarizes them by using the large language model to enhance answer richness.
Tasks
On the AI Search page, click Edit to edit the AI search settings. After you complete the configuration, click OK to apply the configuration changes or click Cancel to discard the configuration changes.
· Search mode
The AI search feature supports the following search modes:
¡ Hybrid Search: Retrieves answers from both the selected knowledge bases and the Internet, and then summarizes and outputs them by using the large language model. This mode is applicable to scenarios that require combining internal data with the latest external information.
¡ Knowledge Base: Retrieves answers only from the selected knowledge bases, and summarizes and outputs answers by using the large language model. This mode is applicable to scenarios involving confidential questions or requiring high reliability for the origin of information.
¡ Online Search: Searches for answers from the Internet, and summarizes and outputs the results by using the large language model. This mode is applicable to scenarios where the involved information requires frequent updates or falls outside the knowledge base scope.
The configurable parameters vary by search mode.
· Model settings
a. From the Model Selection list, select the large language model inference service for the AI search feature to use.
b. (Optional.) For the Diversity parameter, adjust the value by using the slide rail or the input box on the right. Use this parameter to control the randomness or determinism of content generation. A higher value makes the output more random and diverse. A lower value makes the output more focused and deterministic.
c. (Optional.) For the Core Sampling parameter, adjust the value by using the slide rail or the input box on the right. Use this parameter to adjust the determinism of results. A lower value increases determinism, while a higher value introduces more randomness.
· Recommended questions
Recommended questions are questions preset by the AI search feature on the page. You can directly ask these questions.
Parameters
Table 2 Search mode parameters
|
Parameter |
Description |
|
Knowledge Base |
Specify the knowledge bases from which the AI search feature retrieves answers. You can select multiple knowledge bases. |
|
Recalled Entries |
Maximum number of results the inference service returns after it retrieves information from the knowledge bases. |
|
Search Engine |
Specify the search engine used by the AI assistant for Internet searches. You must select a search engine of the online search type that is listed in the Plugins. |
|
Advanced Settings |
Select advanced configuration to set up question understanding, knowledge base threshold, and online search threshold. |
|
Question Understanding |
Whether to perform question understanding on the user's original query. When enabled, the user's original question will be split and expanded. |
|
Knowledge Base Threshold |
Similarity between the returned results and the questions when the AI search feature retrieves information from the knowledge bases. A higher threshold returns more precise results, and a lower threshold returns fuzzier results. |
|
Online Search Threshold |
Similarity between the returned results and the questions when the AI search feature retrieves information from the Internet. A higher threshold returns more precise results, and a lower threshold returns fuzzier results. |
MCPs
About this feature
Model Context Protocol (MCP) is a protocol that describes the identity and capabilities of a model or component in a standard format (such as JSON). Its primary function is to package various external systems into services (MCP services), enabling models to invoke them in a unified manner. For example, through MCP services, AI assistants can call various external tools (such as weather queries and map navigation) to gain the ability to query data or invoke other applications to perform practical operations.
On the MCPs page, the AI assistant, acting as an MCP client, can access the MCP server by creating an MCP service.
Figure 27 MCPs
Tasks
· Add MCP Service
Click Add, set the service name, description, installation method, URL, and authentication information, and then click Test. If the system prompts that the test is successful, you can click OK to complete the creation of the MCP service.
· Modify MCP Service
Click Edit in the Actions column for a MCP service to modify the configuration of the MCP service.
· Delete MCP Service
Click Delete in the Actions column for a MCP service to delete the specified MCP service.
Parameters
Table 3 MCP parameters
|
Parameter |
Description |
|
Service name |
Name of the MCP service. |
|
Description |
Description of the MCP service. |
|
Installation method |
Currently, only the Server-Sent Events (SSE) method is supported. This indicates that the MCP client initiates a request to a specific URL, and the MCP server continuously pushes MCP protocol messages to the client via SSE, enabling remote invocation of this MCP service. |
|
URL |
SSE endpoint address provided by the MCP server. |
|
Header Name |
Name of the authentication header field that the MCP client must include in the HTTP request header. |
|
Bearer Token |
Access token used by the MCP client to access the MCP service. |
Plugins
About this feature
Plugins are components or tools added to an LLM (Large Language Model) to extend its functionality or simplify integration with specific applications. Plugins help users quickly build LLM applications or integrate LLMs into custom programs. Plugins work in conjunction with model services to offer users a wider variety and higher quality of services.
On the Plugins page, you can view and modify plugin information.
Figure 28 Plugins
Tasks
· View plugin information
The Plugins page displays the current plugin information of the AI assistant system in a list format, such as the plugin's name, description, state, and type. The plugin's status indicates its availability on the AI assistant system, including Released and Removed. The Released status indicates that the AI assistant can use the plugin and the Removed status indicates that the AI assistant cannot use the plugin..
· Edit plugin information
Click Edit in the Actions column for a plugin. You can modify the search engine, key, and state of the plugin.
¡ Search Engine: Select the search engine used by the plugin, supported by only some plugins.
¡ Key: Enter the key for the corresponding search engine to establish a connection. Click Connection Test to test the connectivity between the AI assistant and the search engine.
¡ State: Only search engines in listed state can be used by AI search.
System management
Users
About this feature
This feature allows you to manage system users. On the user management page, you can add, edit, and delete users.
Figure 29 Users
Tasks—Search users
· Common search
In the search area on the Users page, enter the login name and phone number, select account status, specify a user creation time range, and then click Search.
To clear all search criteria, click Reset.
· Search users by department
To display all user information for a department, clicking the name of that department on the left side of the user management page.
Tasks—Add a user
1. Click Add.
2. Enter user information, including the username, department, phone number and login name.
3. Click OK.
Figure 30 Add a user
Tasks—Deactivate a user
Use one of the following methods to deactivate users:
· To deactivate a user, click Deactivate in the Account Status column for that user.
· To deactivate users in bulk, select one or multiple users and click Deactivate on the top of the user list.
Tasks—Activate a user
Use one of the following methods to activate deactivated users:
· To activate a user, click Activate in the Account Status column for that user.
· To activate users in bulk, select one or multiple users and click Activate on the top of the user list.
Tasks—Edit a user
1. To edit a user, click Edit the Actions column for that user.
2. Edit user information, including the username and department. The login name cannot be edited.
3. Click OK.
Tasks—Reset a user password
If a user password is forgotten, perform this task to reset the user password.
To reset a user password:
1. Click Reset in the Actions for that user.
2. In the dialog box that opens, enter the new password.
3. Click OK.
Tasks—Delete users
To delete users:
1. Click Delete in the Actions column for a user, or select users and click Delete on the top of the user list.
2. In the dialog box that opens, enter DELETE (case insensitive), and then click OK.
Tasks—Import users
To import users:
1. Click Import on the top of the user list.
2. In the dialog box that opens, click Download Template to download the user information template to the local.
3. Fill in user information in the Excel file as required by the template.
4. Drag the Excel file with user information to the Import User dialog box or click click to upload to upload the user information file.
5. Click OK.
Tasks—Customize columns
To customize columns:
1. Click 
on the user management
page.
2. In the Show/Hide
dialog box that opens, select one or multiple columns or all columns in the Show or Hide list, and then click
and 
to assign columns.
Tasks—Refresh the page
To refresh the user management page, click the
icon on the page.
Roles
About this feature
Role management involves defining a template for a set of system operation authorities, allowing you to specify the menu items a role can access based on service needs. When creating a user, assigning them a role enables batch control of user permissions. This role-based permission management method simplifies permission assignment and enhances system security and manageability.
Figure 31 Roles
Restrictions and guidelines
· Do not add an existing role.
· Do not edit or delete predefined roles such as the super administrator.
· Do not delete a role if a user still belongs to it.
Tasks
· Query roles
Enter the role name or permission character in the text box at the top of the page, select the creation time, and click Search. The matching roles will be displayed. Fuzzy role name search is supported.
To clear all query criteria, click Reset.
To show or hide the search bar, click the
icon on the user management page to show or hide the search icon as
needed.
· Add a role
a. Click Add.
b. Configure the role name, permission character, and set the role order.
c. Set menu permissions for expand/collapse, select all/unselect all, and parent-child linkage.
- Expand/collapse menu permission: By default, menu items are collapsed (only parent items are displayed). After you select the expand/collapse option, all menu items will be displayed.
- Select all/unselect all menu permission: By default, no menu items are selected. After you select the select all/unselect all option, all menu items will be selected.
- Parent-child linkage menu permission: By default, parent and child menu items are not linked. After you select the parent-child linkage option, when a parent menu item is selected, all its child menu items will also be selected.
d. Click OK.
Figure 32 Add a role
· Edit a role
e. Click Edit in the Actions column for the target role, or select the target role and then click Edit on the top of the role list.
f. Edit the role name, permission characters, role order, and menu permission as needed.
g. Click OK.
· Delete a role
After you delete a role, users configured with that role will no longer have the permissions assigned to the role. Be cautious when you delete a role.
You can delete roles in the following methods:
¡ Select one or multiple roles, and then click Delete.
¡ Click Delete in the Actions column for the target role, and then click OK in the dialog box that opens.
· Refresh the page
To refresh the role list, click the 
icon on
the role management page.
Parameters
Table 4 Parameters for adding a role
|
Parameter |
Description |
|
Role Name |
Name of the role, which uniquely identifies a role. |
|
Permission Characters |
Description for the permissions that the role has. |
|
Role Order |
Sort number of the role in the role list. A higher value places the role lower in the list. |
|
Permissions to Menus |
Menu items the role can operate. · Expand/Collapse: Expand or collapse all menu items for easy selection. · Select All/Deselect All: Select all menu items or deselect all of them at a time. · Parent-Child Collaboration: If you select this feature, you will select or deselect a parent menu item and its child menu items. If you do not select this feature, selecting or deselecting a parent menu item does not affect the selection of its child menu items. Currently, this option is selected by default and cannot be canceled. |
|
Remarks |
Remarks for the role. |
Department management
About this feature
Use this page to view and manage department information. On this page, you can add, edit, or delete departments.
Figure 33 Department management
Restrictions and guidelines
The default department is a predefined department. You cannot delete it.
Tasks
· Query departments
Enter the department name in the text box at the top of the page and click <Search>. Tasks that meet the criteria will be displayed below. The department name supports fuzzy search.
To clear the search criteria, click Reset.
To show or hide the search bar, click the
icon on the page to show or hide the search icon as needed.
· Add a department
a. Click Add in the upper-left corner of the page or in the Actions column corresponding to the department.
b. Configure the parent department name, department name, display order, and input information such as the owner, contact number, and email address.
c. Click OK.
Figure 34 Add a department
· Edit department information
d. Select the target department, and then click Edit.
e. Edit the parent department name, department name, display order, owner, contact number, and email address as needed.
f. Click OK.
· Delete a department
Be cautious when you delete a department. Click Delete in the Actions column for the target department, and then click OK in the dialog box that opens.
· Refresh the page
To refresh the department list, click the
icon on the page.
Parameters
Table 5 Parameters for adding a department
|
Parameter |
Description |
|
Superior Department |
Name of the superior department to which the new department belongs. |
|
Department Name |
Name of the department. |
|
Show Sorting |
Sort number of the department in the department list. A higher value moves it further down the list. |
|
Owner |
Name of the department owner. |
|
Contact Number |
Contact number of the department. |
|
|
Email address of the department. |
Menu management
About this feature
Menu management displays the page menu items and their information currently supported by LinSeer Copilot, including:
· SN: The order in which menus appear on pages. A smaller value moves a menu to a higher position.
· Permission ID: Specifies the user permissions required to operate this menu.
· Created At: The time when the menu was created.
Figure 35 Menu management
Tasks
· View menu information
On the menu management page, you can view the menu names, SNs, and permission IDs.
· Edit menu information
Click Edit in the Actions column for the target menu to change its name.
Parameters
About this feature
Perform this task to configure system parameters required for system operation, such as enabling the request URL.
Figure 36 Parameters
Restrictions and guidelines
The Permission page shows system preset parameters. You cannot add or delete parameters.
Each parameter has its specific purpose and application scope.
· For the enableRequestAppKey and enableRequestAppSecret parameters, the provided values are example values and are for reference only. After obtaining the authentication credentials, edit these two parameters based on the actual requirements.
· For other parameters, if you are not certain about their functions or their impact on the system, use the default settings. To avoid errors, do not change the parameter names or key values.
Tasks
· Edit a system parameter
You can modify the parameter name, key value, login verification method, and remarks in either of the following methods:
¡ Click Edit in the Actions column for the target parameter.
¡ Select the target parameter and click Edit above the parameter list.
Figure 37 Edit a system parameter
· Clear cache
Click Clear Cache to delete all cached parameters in Redis.
Parameters
Table 6 Parameters for editing a parameter
|
Parameter |
Description |
|
Name |
Parameter name. Options include: · iFLYTEK voice APPID—Application ID required by AI Assistant to use the iFLYTEK voice input service. · iFLYTEK voice APIKEY—One of the verification credentials required by AI Assistant to use the iFLYTEK voice input service. · iFLYTEK voice APISECRET—One of the verification credentials required by AI Assistant to use the iFLYTEK voice input service. · Linseer hub appKey—One of the verification credentials required by AI Assistant to use the LinSeer large model inference service provided by the LinSeer large model enablement platform. · Linseer hub appSecret—One of the verification credentials required by AI Assistant to use the LinSeer large model inference service provided by the LinSeer large model enablement platform. · Linseer hub grantType—Authorization type obtained from the LinSeer large model enablement platform. The value of this parameter is set to 1. · Linseer hub URI—URL of the API interface provided by LinSeer large model enablement platform. This parameter supports the following key-value formats: ¡ http://<platform_northbound_IP_address>:<platform_API_service_port>/api ¡ https://<platform_northbound_IP_address>:<platform_Web_service_port>/api · Aliyun access key AccessKey—One of the verification credentials for AI Assistant to use the Aliyun verification code login feature. · Aliyun access key AccessKeySecret—One of the verification credentials for AI Assistant to use the Aliyun verification code login feature. · Aliyun signName—SMS signature for users to use a short message sent by Aliyun to log in. The SMS signature identifies the message sender and is used to verify the legitimacy of the message content. · Aliyun template code—SMS content template for users to use a short message sent by Aliyun to log in. · Baidu Translation personal or company account switch—Translation account type used by AI Assistant when translating image generation requests in the LinSeer Multimodal Service application. A value of 0 indicates a company account and a value of 1 indicates a personal account. · Baidu Translation company account APPKEY—One of the verification credentials required for AI Assistant to use a company account to translate image generation requests in the LinSeer Multimodal Service application. · Baidu Translation company account SECREKEY—One of the verification credentials required for AI Assistant to use a company account to translate image generation requests in the LinSeer Multimodal Service application. · Baidu Translation personal account APPID—Service identifier required for AI Assistant to use a personal account to translate image generation requests in the LinSeer Multimodal Service application. · Baidu Translation personal account SECREKEY—Verification credential required for AI Assistant to use a personal account to translate image generation requests in the LinSeer Multimodal Service application. · Shumei APPID—Service identifier required for AI Assistant to use the content compliance verification feature provided by Shumei Technology. · Shumei EVENTID—Event identifier required for AI Assistant to use the content compliance verification feature provided by Shumei Technology. · Shumei SECREKEY—Verification credential required for AI Assistant to use the content compliance verification feature provided by Shumei Technology. · Shumei URL—Service API address of Shumei Technology. · Government WeChat application ID—The service ID required by the AI assistant when integrating with Government WeChat. · Government WeChat application identifier—The identifier of the Government WeChat application. · Government WeChat secret key—The cryptographic key used by the AI assistant for identity authentication when integrating with Government WeChat. · Government WeChat call address—The interface call address for the AI assistant to connect with Government WeChat. · Government WeChat callback address—The callback address for Government WeChat to return data to the AI assistant. · CAS login client address—The client access address required when connecting with the unified authentication system (CAS). · CAS service prefix—The service prefix address required when connecting with the unified authentication system (CAS). · System type for docking—Specifies the type of authentication system to connect with. A value of 0 indicates local login; a value of 2 indicates CAS login. · Knowledge Base Threshold—The threshold for the AI assistant to retrieve the knowledge base, ranging from 0 to 1, used to filter relevant results. · Top-K: Recalled entries in the knowledge base.—The number of results returned by the AI assistant when retrieving the knowledge base, ranging from 1 to 10. · Token expiration time—The session timeout for user login to the AI assistant system, in minutes. · Outside Platform Request Encryption—The encryption key used when connecting with the external platform to ensure data transmission security. · Outside Platform Request Signature—The signature information used for identity verification when connecting with the external platform. · Outside Platform Request URI—The request URI address of the external platform when connecting with it. · ocr service address—Address of the Optical Character Recognition (OCR) service that provides file content parsing. · pdf File Parsing Method: Specifies the parsing method for PDF files. A value of 0 indicates local parsing; a value of 1 indicates OCR parsing. |
|
Key Name |
You cannot edit the key name. |
|
Key Value |
Value of the key. |
|
Login Verification |
Login verification type of the key. Options include: · Logged-in user access—Only login users are allowed. · Non-logged-in user access—Non-login users are also allowed. |
|
Remarks |
Remarks for the key. |
App configuration
About this feature
The AI assistant provides an API for third-party systems to access its resources via application. An application includes an authentication mechanism. When a third-party system uses a specific key to call the LinSeer Copilot's resources, it can only obtain the resources allowed by the application corresponding to the key. These resources are the specified assistants using the big model inference service.
Figure 38 App configuration
Restrictions and guidelines
· The API key and API secret of an application is automatically generated by the system and cannot be configured. The user identifier is used in conjunction with the API key and the API secret to generate a key for the third-party system through an encryption algorithm. This key cannot be edited after you add the application.
· Only invalid applications can be deleted.
Tasks
· Filter applications
In the search area on the top of the page, enter or select search criteria and click Search. The page will then display only the applications that match the criteria. To clear all search criteria and display all applications, click Reset.
· Add an application
To add an application:
a. Click Add.
b. In the dialog box that opens, enter the application name and user identifier. As a best practice, specify the name of the third-party system as the user identifier.
c. (Optional.) Select the check boxes for assistant resources that can be used through the application.
d. Select the validity status of the application in the Status column. When you select Valid, the application has a one-month validity by default. To change the validity period of the application, edit it in the Expiry Date field.
e. (Optional.) Enter a description for the application. The description has only local significance to enhance the application’s identity.
f. Click OK.
Figure 39 Add an application
· Edit an application
To edit an application, click Edit in the Actions column for that application. In the dialog box that opens, edit application settings.
· View the assistants bound to an application
Click Assistant in the Actions column for an application. The dialog box that opens will display the assistants bound to the application, that is, the assistant resources obtained through the application.
In the Application Binding Assistant dialog box, click Copy to obtain the URL for the application. You can use this URL to access the assistant conversation page.
· Delete an application
To delete an invalid application, click Delete in the Actions column for that application.
License information
About this feature
The License Info page primarily displays and manages the system's Licenses. Through this page, users can view the preinstalled licenses and configure connection parameters with the License Server to ensure the device can properly obtain and verify licenses. Additionally, users can review detailed License authorization information, enabling comprehensive monitoring and management of authorization status. This page helps ensure the device operates legally and compliantly, facilitating configuration and maintenance tasks related to License for operations personnel.
Figure 40 License information
Tasks
· View preinstalled licenses
The system automatically applies for preinstalled licenses upon initial installation, allowing users a free one-month trial. Once formal licenses are successfully applied for, the preinstalled licenses will automatically expire.
In the Licensing Method dropdown, select Preinstalled License to view the current preinstalled licenses.
· Connect to the license server
In the Licensing Method dropdown, select Formal License. When establishing the initial connection with the License Server, enter the IP address and port number of the license server, the client name, and the client password, and then click Connect.
· Disconnect
Click Disconnect to disconnect from the license server. The system frees all licenses requested from the license server after disconnection.
· Update licenses
Click Update to refresh the authorization information obtained by the system from the License Server.
· Search for license information
a. In the License Information area, enter the search criteria in the Search bar.
b. Click Search to obtain the matching license information entries.
c. To reset the search criteria, click Reset.
d. (Optional.) To synchronize the license server license information and refresh the list, click Update.
Parameters
Table 7 License server parameters
|
Parameter |
Description |
|
IP Address |
IP address of the license server. |
|
Port Number |
Licensing service port number of the license server. |
|
Client Name |
Name of the client configured on the license server. |
|
Client Password |
Password for the client configured on the license server. |
|
Connection State |
State of the connection to the license server. Options include: · Not Connected. · Connected. · Reconnecting. · Connection Error. · Escape State. In this state, the system can continue to operate for one month. However, if the system fails to reconnect to the License Server during this period, the license will expire. Additionally, if an update authorization operation is performed during the escape period, the license authorization will immediately become invalid. |
Operation log
About this feature
Operation log records the log information generated when the operator executes tasks on the AI assistant.
Figure 41 Operation log
Tasks
· Filter the operation log
In the search area on the top of the page, enter search criteria and click Search. The page will then display only the log messages that match the criteria. To clear all search criteria and display all operation log messages, click Reset.
· View detailed log information
To view detailed information of a log message, click Details in the Actions column for that log message.
· Export operation logs
Click Export, select a time range, and then click OK. The system will export operation logs generated within the specified time range to an excel file.
Login log
About this feature
Login log records information generated when users log in to or log out of the AI assistant.
Figure 42 Login log
Tasks
· Filter login log messages
In the search area on the top of the page, enter search criteria and click Search. The page will then display only the log messages that match the criteria. To clear all search criteria and display all login log messages, click Reset.
· Export operation logs
Click Export, select a time range, and then click OK. The system will export operation logs generated within the specified time range to an excel file.
Scheduled tasks
About this feature
On the Scheduled Tasks page, you can manage all scheduled tasks in the system, such as knowledge base synchronization tasks.
Figure 43 Scheduled tasks
Tasks
· Query scheduled tasks
Enter the task name in the box at the top of the page, select the task group name and task state, and then click <Search>. Matching tasks will be displayed. The task name supports fuzzy search.
To clear the search criteria, click Reset.
To show or hide the search bar, click the
icon to show or hide the search icon as needed.
· Execute a scheduled task once
Click Execute Once in the Actions column for a scheduled task to execute that task once immediately.
· Edit the state of a scheduled task
You can edit the state of a scheduled task in one of the following ways:
¡ Click the slide switch in the Task State column for a scheduled task to start or suspend that task.
¡ Click Edit in the Actions column for a scheduled task. In the Edit Task window that opens, change the value for the Task State field as needed. If you set the task state to Normal, LinSeer Copilot will execute that scheduled task based on the execution expression. If you set the task state to Suspend, that scheduled task will be suspended.
· Edit a task
In the Edit Task window that opens after you click Edit in the Actions column for a scheduled task, you can also edit the execution expression of that task to adjust the scheduled execution time as needed.
The execution expression is a cron expression. The cron expression in a scheduled task contains six fields separated by spaces. Each field represents a time unit.
¡ Field 1: Second field. Set it to a value in the range of 0 to 59.
¡ Field 2: Minute field. Set it to a value in the range of 0 to 59.
¡ Field 3: Hour field. Set it to a value in the range of 0 to 23.
¡ Field 4: Date field. Set it to a value in the range of 1 to 31.
¡ Field 5: Month field. Set it to a value in the range of 1 to 12 or JAN to DEC.
¡ Field 6: Day of the week. Set it to a value in the range of 0 to 6 or SUN to SAT. Use 0 for Sunday.
If the value for a field is a number, the task is executed only when the time unit reaches the specified value. For example, if you set field 1 (second field) to 0, the task is executed at the 0th second of every minute.
For each field, you can use a specific character to add flexibility as follows:
¡ Asterisk (*): Represents any value. For example, if you specify an asterisk (*) in the minute field, the event is triggered every minute.
¡ Question mark (?): Indicates no value is specified. Use this character for the date field and day of week field only to resolve potential conflicts between them.
¡ Hyphen (-): Indicates a range. For example, 10-12 for the hour field represents the time from 10:00 to 12:00.
¡ Comma (,): Indicates additional values. For example, the MON, WED, FRI value for the day of week field represents Monday, Wednesday, and Friday.
¡ Forward slash (/): Indicates an incremental value. For example, the 0/15 value for the minute field indicates that the task is executed every 15 minutes from the 0th minute.
Based on the introduction to cron expressions, the value of 0 0/5 * * * ? in a cron expression indicates that the scheduled task is executed every 5 minutes, starting at the 0th minute and 0th second.
Figure 44 Edit a task
· Refresh the page
To refresh the task list, click the 
icon
on the page.
Data backup
About this feature
Data backup is a key measure to enhance data reliability of AI assistants. Regularly backing up database information allows you to quickly restore the AI assistant to normal during system failures, data corruption, or other unforeseen events.
On the data backup page, you can perform data backup and restoration, and view operation records through the data backup and restoration list.
Figure 45 Data backup
Restrictions and guidelines
During data restoration, any operations performed by users on the system cannot be recorded. To avoid data loss, do not perform any operations during data restoration.
Tasks
· Back up the database
To back up the data in the current database to the server, click Back Up Now.
· Download backup data
In the data backup and restoration list, click Download in the Actions column for a data backup entry to download the corresponding backup data from the server to your local device.
· Restore data from the backup
Click Restore from Backup to upload the previously downloaded backup data and restore the AI assistant to its backed-up status.
· Delete data restoration records and backup data
In the data backup and restoration list, click Delete in the Actions column for a data backup entry to delete the operation record and the backup data saved on the server.
Feedback
About this feature
This page helps you manage issues reported by users. It displays a list of reported issues and lets you view, delete, or export the issues.
Figure 46 Feedback
Tasks
· Query issues
Enter the feedback title in the text box in the top of the page, select a type, and then click Search. Matching results will appear below. The feedback title supports fuzzy search.
To clear the search criteria, click Reset.
· Export feedback entries
Click Export to export the feedback list.
· Delete a feedback entry
Click Delete in the Actions column for a feedback entry. In the confirmation dialog box that opens, click OK.
· View screenshots
Click View Screenshots in the Actions column for a feedback entry to view the pictures uploaded by the user. Before you click View Screenshots, the feedback status displays Unread. After you click View Screenshots, the status changes to Read.
Guidance for using the AI Copilot code plugin
Install the LinSeer Copilot code plugin
Currently, the LinSeer Copilot code plugin can only be installed in VSCode and IntelliJ IDEA. For detailed installation instructions, see H3C LinSeer Copilot Installation and Deployment Guide.
The LinSeer Copilot code plugin has version requirements for VSCode and IntelliJ IDEA. It only supports VSCode 1.70.2 or later and IntelliJ IDEA 2021 or later.
After installing the plugin, you can open the LinSeer Copilot code plugin pane by clicking the plugin icon. All interactions between the user and the LinSeer Copilot code plugin are displayed in the plugin pane.
Auto completion
About this feature
The code plugin can use large model capabilities to help you automatically complete existing code.
Tasks
In the editor pane of VSCode or IntelliJ IDEA, move the cursor to the end of any line of existing code:
· Press <Enter> on the keyboard, and the code plugin will insert a single line of auto-generated completion code.
· Press <Ctrl+Enter> on the keyboard, and the code plugin will insert an entire paragraph of auto-generated completion code.
After the code plugin inserts the code, press the <Tab> key to accept and retain the auto-generated code in the editor pane.
Add comments
About this feature
The code plugin can help you understand code meaning and generate corresponding comments for the code.
Tasks
In the editor pane of VSCode or IntelliJ IDEA, select a block of code, right-click, and select H3C LinSeer Copilot-Add Comments from the menu bar. The LinSeer Copilot code plugin pane will display the code with added comments. You can copy the commented code or directly insert it into the editor pane.
Explain code
About this feature
The code plugin can help you parse the content and functionality of the code in detail and then informs you through text in the plugin pane.
Tasks
In the editor pane of VSCode or IntelliJ IDEA, select a block of code, right-click, and select H3C LinSeer Copilot-Explain Code from the menu bar. The LinSeer Copilot code plugin pane will display the parsed content of the code. The parsed content can be copied using the icon in the upper-right corner of the text.
Custom queries
About this feature
In the plugin pane, you can also engage in intelligent question answering (Q&A) with LinSeer Copilot through prompts. The types of Q&A include:
· Direct Q&A without context.
· Q&A based on selected code content. This method is more flexible than adding comments and explaining code and can be used for scenarios such as troubleshooting, parsing, and optimizing code.
Tasks
· Direct Q&A without context
In the Q&A box of the plugin pane in VSCode or IntelliJ IDEA, enter a prompt directly and click <Send> to interact with LinSeer Copilot.
· Q&A based on selected code content
a. In the editor pane of VSCode or IntelliJ IDEA, select a block of code, right-click, and select H3C LinSeer Copilot-Custom Query.
b. In the Q&A box of the plugin pane, enter a prompt based on the code content and click <Send> to interact with LinSeer Copilot about the selected code.
Configure the LinSeer Copilot code plugin
· In the editor pane of IntelliJ IDEA, right-click and select H3C LinSeer Copilot-Settings from the menu bar to modify the large model used by the plugin.
· In the plugin pane of VSCode or IntelliJ IDEA,
click the 
icon at the top to create a new conversation. Click the delete icon
in the conversation dropdown to remove a specific conversation.
· In the plugin pane of VSCode or IntelliJ IDEA,
click the 
icon at the top to set the conversation mode during interaction
with LinSeer Copilot:
¡ Single-turn conversation: LinSeer Copilot answers directly based on the user's question without considering context.
¡ Multi-turn conversation: LinSeer Copilot uses the most recent Q&A pairs as context to answer questions.
· In the plugin pane of VSCode or IntelliJ IDEA,
click the 
icon at the top to view all conversation records saved as .json
files.
· In the plugin pane of VSCode, click the 
icon
at the top to refresh the conversation UI interface.