Login System

New users of ScienceOne need to register and log in to their personal account.

1.1 Access

Official ScienceOne website: https://www.scienceone.ai.

Google Chrome or Microsoft Edge browser is recommended.

1.2 Register

When accessing ScienceOne for the first time, you need to register an account. The system provides phone number and email registration options for individual users. The page defaults to phone number registration; click Email at the bottom to switch to email registration. Enter your phone number or email address in the input field and click Send to receive a verification code. Enter the correct verification code to complete registration and automatically create your account.

1.3 Login

Two main login methods are supported:

• Passcode: Enter your phone number and click Send to receive a verification code. Unregistered phone numbers will have an account auto-created upon verification.
• Password: Enter your phone number and account password to log in.

Additionally, you can log in via Email, Account, or Organization using the options at the bottom of the page.

Research

2.1 Conversation Modes

The Research provides multiple usage modes for scenarios including literature research, subject questions answering, research report writing, visualization report generation, and literature reviews.

Deep Research: More comprehensive, suitable for research, reports, mining research ideas scenarios, integrating massive information to fully support research decision-making.

Literature Review : More transparent, specialized for ultra-long literature review scenarios, deeply analyzing tens of thousands of papers, comprehensively organizing research contexts.

Innovation Evaluation: More rigorous, suitable for self-inspection and auxiliary peer review of projects and papers before submission. It accurately analyzes areas for improvement to refine manuscripts for higher review standards.

Academic PPT: More professional, suitable for proposal reports, academic conference presentations, research result display scenarios, upload papers to generate with traceable images and charts.

Scientific Illustration: More precise, suitable for paper chart drawing, report illustration design, academic presentation visualization material production scenarios, precisely matching research content.

Academic Poster: More intuitive, suitable for academic exchanges and result display scenarios, upload papers to generate with visualized research highlights.

MatScientist: Focuses on material characterization analysis, structural and property research, and assists phase identification, performance interpretation and material property trend prediction.

MolChemist: Covers molecular structure, physicochemical properties, toxicology and metabolomics data mining, supporting drug screening, compound evaluation and molecular knowledge discovery.

MolSimulator: Adapts to thermodynamic calculation, molecular/protein structure simulation and parameter fitting, and delivers intelligent interpretation for stability screening and computational outputs.

BioResearcher: Tailored for sequencing, gene editing, cell and synthetic biology experiments, providing one-stop services including experimental protocol design, omics data processing and result analysis.

2.1.1 Deep Research

Deep Research is currently the default conversation.In this mode, a single-agent closed-loop (AgentLoop) architecture is adopted. The agent thinks about your question and outputs an answer.

The system consists of two parts: Process (left) and Workspace (right). The left Process Area displays the current execution process, including thinking process, execution details, and final generated results. The right Workspace contains three sections: "Tasks" shows output results of agent execution steps; "Information" summarizes information collected through skills, web sources, and literature databases; "Files" displays all files generated for the current task.

2.1.2 Deep Survey

Select the "Deep Survey" button below the homepage dialog box to use Literature Review Mode.

In this mode, you can input your research questions to obtain professional reviews of tens of thousands of words. Before starting the task, you can select literature libraries, online sources, and target language.

After the task starts, the agent extracts detailed information such as titles and keywords of the review, generates a modifiable review outline, and waits for your confirmation before continuing execution.

After task completion, the system displays the complete the literature review report including abstract, main text, and references. Click "Download" to download the full report in PDF, Word, or MD format.

2.1.3 Innovation Evaluation

At the bottom of the homepage dialog box, select the “Innovation Evaluation” button to enter the review report generation mode. Users can click the upload area to open the local file selector, or directly drag and drop a local file for upload. The system supports uploading one PDF, Word, or Excel file as the material to be reviewed, with a maximum file size of 20 MB.

Before starting the task, users can select the appropriate review guidelines or evaluation criteria and choose whether to enable or disable information sources as needed. Based on the selected criteria and information source settings, the system will parse, match, and evaluate the uploaded materials, and generate the corresponding review results and report.

After the user uploads the materials and selects the corresponding review guidelines or criteria, the system parses both the review materials and the selected standards to identify the core information, evaluation dimensions, and judgment basis of the review object. On this basis, the system automatically plans search tasks according to the review topic and standards, and supplements the evaluation with domain knowledge relevant to the current review task.

During the domain knowledge construction process, the system performs relevance filtering, key information extraction, and multi-round verification of the retrieved information to ensure the effectiveness and reliability of the knowledge introduced. It then conducts a joint analysis of the review content, review guidelines, and constructed domain knowledge to form structured evaluation results, and outputs a review report containing evaluation conclusions, analytical basis, and cited evidence. Reports can be generated in both HTML and Markdown formats for convenient display, archiving, and subsequent review.

2.1.4 Academic PPT

At the bottom of the homepage dialog box, select the “Academic PPT” button to enter the academic PPT generation mode. It supports uploading 1 academic document in PDF, DOC, or DOCX format with a file size within 20MB by clicking the attachment. Before uploading, you can choose the output language (Chinese/English, default is English) and the theme color, and the Agent will create the academic PPT based on your selected configuration.

The Academic PPT Agent will invoke the ppt-generator skill, which is divided into four phases. First, it will perform input document confirmation and source document parsing. The Agent will read the source file and parse the paper content and chart information, and during the parsing process, detailed processing logs will be displayed in the workspace. Second, based on the configuration you selected, the Agent will generate recommended parameters for the academic PPT product for you to confirm or modify. In the second phase, a presentation outline will be generated. The Agent will generate and display a complete PPT outline structure, and after you confirm or modify the outline, you will proceed to the next phase.In the third stage, PPT content is generated page by page, and the right workspace will display the progress of PPT generation for each page. In the final stage, PPT rendering is completed and a downloadable PPT document is provided.

Throughout the entire generation process, the right-side workspace will display two main content modules in real time. The Task Module shows detailed processing logs, including records of the entire process of file reading, content parsing, outline generation, and PPT production. The File Module displays files generated during the task in folders, including charts extracted after academic literature parsing, txt and json files of processing logs, markdown result files from each generation stage, PPT files generated page by page, etc.

2.1.5 Scientific Illustration

At the bottom of the homepage dialog box, select the “Scientific Illustration” button to enter the scientific illustration generation mode. The underlying model of the scientific illustration intelligent agent uses the MultiModal Machine Learning inference model S1-Omni-Image, following the "Think before Generate" paradigm and supporting two core functions: image generation and image editing. Trained on massive scientific illustrations covering six major disciplines （mathematics, physics, chemistry, astronomy, geography, and biology）.The model is compatible with common types of scientific diagrams including flowcharts, architecture diagrams, and principle diagrams. It visualizes the full drawing reasoning process in real time during generation, enabling one-stop creation and refined editing of scientific illustrations.

You can describe the content of the required scientific research illustration in natural language. After submitting the illustration request, the system will automatically execute two key steps to complete the illustration generation. First, it will understand the scientific illustration requirements and conduct a relevance assessment. The Agent will analyze your description, determine whether it meets the requirements for scientific illustrations, and evaluate the security and compliance of the content. After passing the assessment, the Agent will call the S1-Omni-Image model to draw the scientific illustration, generate professional academic-style images based on your requirements, and add an AI-generated watermark to identify the source. After the illustration generation is completed, the Agent will display the results of the scientific research illustration generation.Supports using the like/dislike feature to provide feedback on the generated results, helping the Agent continuously optimize.

Additionally, the Agent supports multi-round conversations to refine the illustration effect. You can also continue to provide modification suggestions or new illustration requirements in the current conversation until your academic needs are met.

2.1.6 Academic Poster

At the bottom of the homepage dialog box, select the “Academic Poster” button to enter the academic poster generation mode. It supports uploading 1 academic document in PDF, DOC, or DOCX format with a file size within 20MB by clicking the attachment. Before uploading, you can select the output language (Chinese/English, default is English), size, and theme color, and the Agent will create the academic poster based on your selected configuration.

The academic poster agent will invoke the poster-generator skill to gradually complete the academic poster generation task. Through eight steps of parsing the document -> confirming parameters -> content curation -> color design -> title design -> layout optimization -> font design -> rendering the poster -> final verification, a downloadable academic poster will be generated.

Throughout the entire generation process, the right-side workspace will display two main content modules in real time. The Task Module shows detailed processing logs, including records of the entire process of file reading, content parsing, and poster creation. The File Module displays files generated during the task in folders, including charts extracted after academic literature parsing, txt and json files of processing logs, markdown result files of the eight steps, the final academic poster PPT file, etc.

2.1.7 MatScientist

The MatScientist is designed for the interpretation of characterization data such as XRD, NMR, and Raman, as well as phase identification. It supports analysis of material structure, electrochemical performance, stability, and structure–activity relationships, assisting in the understanding and prediction of material properties.

Select the "MatScientist" button below the dialog box on the homepage to enter the MatScientist. You can upload up to 10 files of various formats, each within 20 MB, by clicking the attachment icon. Before starting a task, you may select the required Skills and MCPs, and enable or disable the knowledge base and information sources as needed.

2.1.8 MolChemist

The MolChemist is designed for the retrieval, analysis, and knowledge discovery of molecular structures, physicochemical properties, biological activities, ADMET/toxicity predictions, and metabolomics data, assisting in drug research and compound evaluation.

Select the "MolChemist" button below the dialog box on the homepage to enter the MolChemist. You can upload up to 10 files of various formats, each within 20 MB, by clicking the attachment icon. Before starting a task, you may select the required Skills and MCPs, and enable or disable the knowledge base and information sources as needed.

2.1.9 MolSimulator

The MolSimulator supports protein assembly interface analysis, material thermodynamic stability screening, equation-of-state fitting, and primer/structure parameter calculations, assisting in thermodynamic property analysis and the interpretation of computational results.

Select the "MolSimulator" button below the dialog box on the homepage to enter the MolSimulator. You can upload up to 10 files of various formats, each within 20 MB, by clicking the attachment icon. Before starting a task, you may select the required Skills and MCPs, and enable or disable the knowledge base and information sources as needed.

2.1.10 BioResearcher

The BioResearcher is designed for tasks such as gene editing, omics sequencing, genetic analysis, flow cytometry, synthetic biology, and animal experiments, covering scenarios from molecular biology to applied life sciences. It supports experimental design, data processing, and result interpretation.

Select the "BioResearcher" button below the dialog box on the homepage to enter the BioResearcher. You can upload up to 10 files of various formats, each within 20 MB, by clicking the attachment icon. Before starting a task, you may select the required Skills and MCPs, and enable or disable the knowledge base and information sources as needed.

2.2 Attachments and Settings

2.2.1 Upload Attachments

Supports uploading mainstream document formats: pdf/pptx/xls/xlsx/csv/docx/md/txt/epub/ofd/wps

More file formats are continuously being expanded. Supports uploading up to 10 files at once, each not exceeding 20MB.

2.2.2 Set Skill, MCP, Library, and Online Sources

Click Settings to select the agent's skills, MCP servers, information sources, and output configuration. Information sources include Library and Online sources. Output configuration options vary by mode — for example, in Literature Review mode, you can set the target language; in Academic Poster mode, you can set both the target language and theme color.

• Skills:

Displays the skills available for invocation. You can choose which skills to call based on your specific research scenario.

• Click the Skill button in the input bar to open the skill list. You can click a skill to load it directly into the input box for invocation.
• Click Manage My Skills to open a panel where you can select up to 10 skills for the agent to invoke automatically as needed.
• Skills with the toggle enabled will be triggered automatically on demand.
• Skills with the toggle disabled can still be loaded into the input box manually, forcing a 100% explicit invocation in the format /[skill-slug].
• If you try to enable more than 10 skills, a message will appear: "The maximum number of auto-triggered Skills has been reached."
• MCP

Displays your enabled MCP servers. You can configure external MCP servers independently, giving the research assistant continuously expandable tool-dispatching capabilities.

• Click the MCP button in the input bar to view the list of Enabled MCP Servers.
• Click Manage My MCP Servers to open a panel where you can select up to 10 MCP servers for the agent to invoke automatically on demand.
• The first time you enable an MCP server, you will need to enter authentication credentials. The platform securely encrypts and stores your credentials, so you will not need to re-enter them for the same MCP server in the future.
• While entering authentication credentials, you can click Test Connection to verify that the MCP server is connected correctly.
• Library: Library databases are created by you according to research needs for storing selected papers. You can select which literature databases\libraries the agent is allowed to access.
• Online Sources: Online sources include academic and Web search. After enabling Academic, the agent connects to academic databases including 170 million scientific papers from CAS Literature Information Center, 2.4 million arXiv preprints, and global paper resources from Google Scholar;After enabling Web search, the agent actively searches the internet for relevant materials.

2.3 Conversation

2.3.1 Conversation Termination

During any conversation process, click the terminate button to immediately stop the ongoing conversation.

2.3.2 New Conversation

After the current conversation ends, click "New Chat" to start a new round of conversation. The new conversation is not affected by other conversation contexts.

2.4 Progress Viewing

In Research Mode, tasks are divided into multiple steps. Click expand to view all task progress. Also, click the expand button on the right side of each step in the process to view detailed process of each step. Click specific execution actions to view results of that action in the workspace.

2.5 Workspace

2.5.1 Progress Navigation

Complete task progress is displayed as a progress bar. Users can click "Left/Right" switch buttons to navigate between steps, or drag the progress bar to control displayed steps.

2.5.2 Search

Search results for each step are displayed in Workspace. Click to view original link content.

All search results for the task can be viewed in Information in the workspace.

2.5.3 Generated Files

Process files such as analysis results may be generated during execution. Click specific actions to view file content in the workspace, with support for switching between source code and preview views.

All task process results can be viewed by clicking "File".

All task final results (such as visualization reports, reviews, etc.) can be viewed by clicking the "File" after task completion or finding the final result file in the workspace.

Library

Library provides in-depth interpretation services for carefully selected papers for researchers. Integration with literature management tools such as Zotero will be progressively implemented to further facilitate researchers' daily work.

3.1 Create A Library

When creating a literature library, select the corresponding library type based on literature type. The system has designed specialized interpretation logic for different literature types. Selecting the appropriate library type yields better usage results.

3.2 Upload Literature

File upload supports two methods.

• Direct Upload: Batch upload literature files directly on the literature library management page, suitable for initial library creation or centralized literature organization.
• Add Anytime: When using the literature library for research, you can add new literature files anytime as needed, facilitating timely supplementation of relevant materials. Both methods support mainstream document formats. Choose the appropriate upload method based on your specific usage scenario.

3.3 View Library

All literature libraries you create are visible only to you by default, ensuring privacy and security of research materials.

To collaborate with partners, you can add collaborators through the "Member Management" function and set access permissions for them.

Permissions are divided into three levels: "Administrator" has full operation permissions; "Collaborator" can create directories and upload literature; "Member" can only view, and cannot modify content. As the literature library creator, you can adjust member roles or remove members at any time. To ensure literature library security, team member management permissions are limited to the creator only.

3.4 Use Library

3.4.1 Conversation with Total Library

You can converse with all papers in the literature library, or select a single paper for intensive reading.

3.4.2 Conversation with One Paper

In a conversation with single paper, you can deeply interpret the selected paper, or switch to full library conversation mode at any time.

3.4.3 Guide for Paper

The Guide provides core content analysis of papers, covering research conclusions, research objectives, research methods, rationality, limitations, core viewpoints, evidence support, application prospects, and impact.

3.4.4 Research Graph for Paper

Research Graph provides citation tracing functionality, taking the current paper as the center to trace its references while exploring the paper's citation situation, helping you understand the complete academic context.

The visual network intuitively displays citation relationships between papers, helping you clearly see inter-paper citations for quick identification of key research gaps.

3.4.5 Chat for Paper

The system pre-interprets each paper, generates possible frequently-asked questions and pushes them to the interface. You can directly click to start a conversation.

You can also ask questions independently based on actual needs. The system provides targeted paper interpretation and reference sources. Image content from papers is displayed directly in answers.

3.4.6 Notes for Paper

The library provides note functionality for recording reading notes. After switching the right sidebar to the "Note" module, select text on the left to automatically create annotations, then add your note content.

Feed

Click on "Feed" in the left navigation bar to enter the module. The Feed module aggregates over 1 billion academic information sources, bringing together interdisciplinary papers, news and offers exclusive topic subscription features and automated task push for research briefings to help you efficiently track the latest developments in your research field.

4.1 News Browsing

Scientific research news is arranged in a waterfall flow in the Recommend interface in the form of cards, divided into three types of cards: papers, news, and topics. Each card contains key information such as source channels (e. g., "ScienceNet", "arXiv"), title, author, abstract, publication time, accompanying images, click-through rate, etc. After entering the module, the system defaults to displaying the latest information of all types, and scrolling down will load more content.

4.1.1 News Screening

The system provides two filtering methods, subject classification and keyword filtering, to help you quickly locate the information you need:

• Subject Classification Filtering: Feed currently covers information on a total of ten major disciplines, including Mathematics & Physics, Chemistry, Life Science, Earth Science, Engineering and Materials, Information Science, Management Science, Medicine, Social Science, and Humanities. Each discipline has subordinate secondary disciplines. You can directly select the discipline of your research field or other disciplines of interest to quickly browse recent scientific research trends.
• Keyword Filtering: Enter the keyword of the information you want to locate in the input box in the upper right corner, and scientific research news will be presented below according to content relevance.

4.1.2 Paper

• Card Operation Area: For cards of the paper type, click on the top right corner of the card to open the more operations menu, where you can select [Open in Research] - directly access Deep Research, Academic PPT, and Academic Poster agents; select [View PDF] to directly view the original paper; select [Add to Library] to save the paper PDF to your personal document library.

• Paper Details Page: Click on the paper card to open the details drawer on the right, where you can view key information such as the paper's author, abstract, DOI number, publication date, etc. in detail. You can click the [AI Summary] button, and the system will interpret information such as the problem the paper aims to solve, related research, and main content in the form of Q&A. Different papers also have dedicated answers to questions about that paper, helping you accurately and efficiently understand the research content of the paper without having to read the original text.

• Paper Sharing: Click the [Share] button below, and it will generate a paper sharing poster and link for you. You can copy the sharing link and share it with your research team to jointly read the latest papers, discuss the key points summarized by AI, and help the research team stay updated on research trends in real time.

4.1.3 News

Feed aggregates multi-source news, from government agencies, university research institute laboratories to science and technology public accounts, news media, etc. All the information you care about can be found here. Click on the news card to quickly jump to the original website. Multiple sources such as WeChat public accounts, websites, and videos help you stay informed about the latest scientific research policies, conference information, major breakthroughs, and other scientific research news.

4.1.4 Topic

Feed has a large number of officially recommended topics built-in, aggregates high-quality news sources, and continuously updates information related to the topic under the source. You can search for topics in the fields and directions you are interested in to quickly view information in that area, or click the [Subscribe to Topic] button to continuously receive updated information on this topic.

The special topic also supports the sharing function. Click the [Share] button below, and it will generate a sharing poster and link for this topic. You can copy the sharing link and share it with your research team to read the special topic together, helping the research team receive real-time research trends.

4.2 Topic Subscription

The subscription section has opened the special topic subscription function. The subscription list will include the official special topics you have subscribed to and the self-created subscription topics. The number of updates is displayed in a blue corner mark in the upper right corner of the special topic card, helping you intuitively understand the information update status of the topic. Click on a special topic in the list to view the continuously updated included information of that topic.

4.2.1 Self-built Topic Subscription

• Create a Special Topic Subscription: By creating your own subscription topics, you can track in real-time the scientific frontier information you are more concerned about. Click the 【Subscribe to New Topic】 button, enter the topic name and description, customize the tracking targets from over 1 billion information sources across the web, and a single topic can subscribe to up to 50 sources at most. Adding information sources supports keyword-based targeted search via the search box + category tag selection.
• Edit: Topics that have been successfully subscribed support modifying the tracked information sources. You can add new information sources or delete the added ones. The new information source settings will take effect immediately. Information from newly added information sources will be updated to the details of this special topic thereafter.

4.2.2 Automation Tasks

Based on the thematic content, you can create automation tasks, enabling the Feed Agent to assist you in continuously tracking, organizing scientific research briefings, and pushing them to your WeChat and Feishu channels. Scientific frontier information will be actively aggregated for you, returning the time spent on tracking information to research itself.

There is no upper limit to the number of automated tasks that can be set for each topic, but it is recommended to schedule multiple tasks at different time points, as tasks scheduled at the same time will cause queuing, and you will need to wait for the Feed Agent to complete one automated task before starting the next.

• Create an automation task: Find the topic you want to track in the subscribed topic list, select the topic, click [New Auto Task], and an automation task based on the updated information of this topic will be created.Enter your requirements, which can be the latest developments of an enterprise or a specific research area you are interested in, such as natural language processing (NLP). You can ask it to filter recent technological advancements, industry progress, media discussions, open-source code, and other specific content that interests you. Finally, you can request it to generate research briefs in formats such as WORD, PDF, HTML, or PPT. The Feed Agent will organize the research briefs according to your requirements, start the task at the execution time you set, and push it to the push channel you set after the task is completed.
• Channel Connection: You can follow the instructions in the Channel Management section of the User Center to connect your WeChat and Lark to the Feed Agent. When setting up automated tasks, check the connected channels, and the Agent will push scientific research briefings to the specified connected channels.

Click on an already created automation task in the automation task list of the topic, and the details of that automation task will pop up on the right. You can immediately test whether the automation task is executed normally, pause and resume the task, edit the task, delete the task, and view the details of the operation log here.

• Run Now: Click the 【Run Now】 button, and the Feed Agent will immediately start executing this task. You can test here whether the channel access is normal and whether the content of the scientific research briefing push meets your expectations, and make timely adjustments to find the content of the scientific research briefing that satisfies you more.
• Pause and Resume Task: If you temporarily do not want to receive the push notifications for this automation task, simply click the [Pause Task] button to pause it; after pausing, you can also click the [Resume Task] button to restart the execution of the automation task.
• Edit Task: The created task supports re-editing, and after confirmation and saving, it will take effect the next time the task is executed.
• Delete Task: You can delete created tasks, but note that task records will not be saved, while previously generated scientific research briefings will remain in your push channel.
• View Run Log: The list below presents the details of the run log for this automated task, recording the task's start time, duration, messages, Tool calls, and status.

Due to the special restrictions of WeChat Channels, please interact with him frequently to successfully receive the research briefings.

[还要一张拼接的结果图]

ScienceHub

5.1 Model

The Model module serves researchers across all disciplines. It integrates domain-specific large scientific models and agent resources available on the platform, covering a range of cutting-edge research fields including solar physics, materials innovation, carbon accounting, particle physics, and near-space science. Users can browse model information and access available capabilities for online trial.

5.1.1 Accessing the Model Entry

• Click Model in the left navigation bar to enter the module.
• The "All" models page is displayed by default.

5.1.2 Browsing and Searching Models

• Category switching: Three tabs are available at the top: All, Model, and Agent
• Model — Displays all models of the Model type
• Agent — Displays all models of the Agent type
• Search: Enter keywords (both Chinese and English supported) in the search box, and the system will perform fuzzy matching against model names and descriptions.
• Sorting: Results are sorted by Popular by default. Click the arrow to toggle between ascending and descending order.

5.1.3 Viewing Model Details

• Click any model card to open its detail page.
• The detail page presents information such as research background, core capabilities, and key outcomes. Details may vary slightly from model to model.
• Model example — ScienceOne · MatMind:
• On the model list page, click the ScienceOne · MatMind Agent card to access its dedicated detail page.
• A Chinese/English language toggle is available at the top right of the detail page to fit different reading preferences. Users can scroll down to browse content section by section, or click the Explore In-depth button to jump directly to the full introduction covering research background, core capabilities, key outcomes and more.
• Click the Try Now button on the page to redirect to the MatMind Intelligent Materials Innovation Platform. The platform is equipped with multiple specialized agents for niche research directions, covering functional scenarios such as material solution generation, material design, and theoretical material computation. Select the corresponding agent to start using.

5.2 Dataset

The Dataset module lets you browse and search authoritative scientific datasets. You can bookmark any dataset you are interested in, turning scattered data access into a centralized, ready-to-use experience.

5.2.1 Finding the Dataset Entry

• Click Dataset in the left navigation bar.
• The default view shows Public Datasets.

5.2.2 Browsing and Searching Datasets

• Switch tabs: Two tabs appear at the top:
• Public Datasets — displays all datasets on the platform.
• My Favorites — displays datasets you have bookmarked.
• Search: Enter keywords (Chinese or English) in the search box; the system performs fuzzy matching against names and descriptions.
• Discipline filter: Use the top dropdown to filter by discipline (e.g., Life Sciences, Pharmaceutical Sciences, Chemistry). All disciplines are shown by default.
• Sort: Defaults to Popular order; you can switch to Name order and click the arrow to toggle ascending/descending.
• Count: The title area displays the total number of invocable Datasets, updating live as you switch tabs.

5.2.3 Adding Datasets to Favorites

• Hover over a dataset card — a ··· (More) button will appear.
• Click ··· to expand the options: Marked / Unmarked and Share.
• Click Add to Favorites → the button changes to Added to Favorites, and the dataset is synced to My Favorites.
• Click Added to Favorites again to remove it from favorites.
• Note: After removing a favorite, the page does not refresh automatically — you need to refresh manually to see the change.

5.2.4 Viewing Dataset Details

• Click any dataset card to open its detail page.
• The detail page shows: dataset name, all discipline tags, data source, license, and access method.
• Three action buttons are provided:
• Visit Official Website — jump directly to the dataset's official source page.
• Marked / Unmarked — bookmark or unbookmark.
• Share — copy the detail page link to share useful datasets.
• The Overview section at the bottom displays the full documentation for the dataset.

5.3 MCP

The MCP module is used to connect and manage external MCP tools. They can be invoked on demand during scientific conversations, helping researchers flexibly build their own intelligent toolkits.

5.3.1 Finding the MCP Entry

• Click MCP in the left navigation bar.
• The default view shows Public MCP Servers.

5.3.2 Browsing and Searching MCP Tools

• Switch tabs: Three tabs appear at the top:
• Public MCP Servers — displays all tools on the platform.
• My Favorites — displays tools you have bookmarked.
• My MCP Servers — displays tools you have created.
• Search: Enter keywords (Chinese or English) in the search box; the system performs fuzzy matching against names and descriptions.
• Discipline filter: Use the top dropdown to filter by discipline. All disciplines are shown by default.
• Sort: Defaults to Popular order; you can switch to Time or Name order and click the arrow to toggle ascending/descending.
• Count: The title area displays the total number of invocable MCPs, updating live as you switch tabs.

5.3.3 Adding MCP Tools to Favorites

• Hover over a tool card — a ··· button will appear.
• Click ··· to expand: Marked / Unmarked and Share.
• Under Public MCP Servers, click Marked → the button changes to Unmarked, and the tool is synced to My Favorites.
• Click Marked again to remove it.
• Note: After removing a favorite, the page does not refresh automatically — you need to refresh manually.

5.3.4 Viewing MCP Tool Details

• Click any tool card to open its detail page.
• The detail page shows: tool name, discipline tags, developer, creation time, and popularity.
• Available actions on the detail page:
• Marked / Unmarked
• Share — copy the detail page link to share useful tools.
• The Overview section displays the tool's full documentation.
• The Tool Testing section — invoke the tool directly to test it and view returned results.

5.3.5 Creating Your Own MCP Tool

• On the MCP page, click the + create button in the top-right corner.
• Fill in the form:
• Name (required)
• Description (required) — a brief explanation of the tool's purpose.
• Type: Select SSE or Streamable HTTP.
• URL: Enter the tool's access URL (must start with http:// or https://).
• Authentication method: Choose one of three options:
• No Auth — connect directly.
• Bearer Token — provide an access token.
• Custom Headers — provide key-value pairs.
• Discipline: Select discipline type
• Tags: Choose recommended tags or type custom ones; press Enter to confirm, Backspace to delete; up to three tags are supported.
• Documentation: Click or drag to upload a markdown file
• Click Confirm.
• After successful creation, the new tool appears at the top of My MCP Servers.
• You may also click Test Connection first — if the connection passes, the MCP is functioning normally; then click Confirm.

5.3.6 Managing My MCP Tools

• In My MCP Servers, you can Edit or Delete any tool.
• The tool name cannot be changed after creation; all other fields are editable.
• Deletion requires a second confirmation.

5.4 Skill

The Skill module manages methods and workflows that AI can invoke. The platform includes thousands of high-quality built-in Skills covering common research scenarios, allowing users to turn their unique research experience into reusable assets.

5.4.1 Finding the Skill Entry

• Click Skill in the left navigation bar.
• The default view shows Public Skills.

5.4.2 Browsing and Searching Skills

• Switch tabs: Three tabs appear at the top: Public Skills, My Favorites, My Skills.
• Search: Enter keywords (Chinese or English); the system performs fuzzy matching.
• Discipline filter: Use the top dropdown to filter by discipline. All disciplines are shown by default.
• Sort: Defaults to Popular order; you can switch to Time or Name order and toggle ascending/descending.
• Count: The title area displays the total number of invocable Skills, updating live as you switch tabs.

5.4.3 Adding Skills to Favorites

• Hover over a skill card — a ··· button will appear.
• Click ··· to expand: Marked / Unmarked and Share.
• Under Public Skills, click Marked → the button changes to Unmarked, and the skill is synced to My Favorites.
• Click Marked again to remove it.
• Note: After removing a favorite, the page does not refresh automatically — you need to refresh manually.

5.4.4 Creating Your Own Skill

Method 1: Create via Research

• Hover over the Create button.
• Two options appear. Click Create via Research.
• You are redirected to the Research, which auto-calls the /skill-creator skill, which creates the skill based on your description.

Method 2: Create via File Upload

• On the Skill page, click the + create button.
• Drag and drop, or click to upload a skill folder or .zip archive.
• The system auto-detects and pre-fills: skill name, description, and version number (defaults to 1.0.0 if not specified).
• Fill in the Slug — only hyphens, numbers, and lowercase letters are allowed; this cannot be changed after creation.
• Discipline: Select from the dropdown.
• Tags: Choose recommended tags or type custom ones; press Enter to confirm, Backspace to delete; up to three tags supported.
• Click Confirm; the system automatically reviews the submission (prohibited content, file integrity, code security, etc.).
• If review fails, the specific reason is shown (e.g., prohibited content detected, missing SKILL.md, etc.).

5.4.5 Publishing a Skill Publicly

• In My Skills, find the skill you want to publish.
• Click the Publish button.
• A second confirmation dialog appears: "Once published, this Skill cannot be unpublished."
• After confirming, the skill appears in Public Skills and is marked as Published.

5.4.6 Updating a Skill

Method 1: Update via Research

• Update the skill through conversation.
• After the conversation ends, a Save button appears.
• Click it to update the current skill.

Method 2: Update via File Upload

• In My Skills, click the Update button.
• Re-upload the skill folder or .zip archive.
• You may revise the name, description, version number, discipline, and tags. The version number must be incremented from the current version.
• Click Confirm Update; the system re-runs automatic review.

5.4.7 Deleting a Skill

• In My Skills, click the Delete button → confirm a second time to delete.
• Note:
• Skills that have not been published are permanently removed.
• Skills that have been published disappear from My Skills but remain visible in Public Skills.

5.4.8 Viewing Skill Details

• Click any skill card to open its detail page.
• The detail page shows: skill name, discipline tags, developer, version number, creation time, and popularity.
• Available actions:
• Marked / Unmarked
• Share — copy the detail page link.
• The Overview section displays the skill's full documentation.
• The Installation section provides installation instructions.

5.4.9 Using Platform Skills on External Platforms

• On the skill detail page, find the Installation section.
• Three cross-platform installation methods are provided:
• Chat Installation — copy the prompt and send it to any AI agent.
• CLI Installation — for agents that support the Skillhub CLI.
• ZIP Installation — download the skill archive and upload it to any AI agent.
• The Skill Files panel on the right side of the detail page provides a file tree directory with support for search, downloading the archive, and copying individual file contents.

5.4.10 Using a Skill

• Hover over a skill card — a ··· button appears.
• Click ··· to expand the options: Use, Marked / Unmarked, and Share.
• Under Public Skills, click Use → automatically jumps to the Research, which displays how to use the skill in the conversation.

History

The History module records all your research tasks in the Research, helping efficiently manage and track research progress. Click any task to enter the detailed interface and view complete information and latest progress. Support real-time monitoring of task progress status, and click to view details.

6.1 Records

Clearly displays all created research tasks, including task names, creation times, last update times, and brief descriptions. You can filter historical conversation tasks by different modes.

History records display current conversation task status (in progress, manually stopped, execution failed). You can view different types of historical tasks. Click any history record to enter that conversation and view current progress and related data.

6.2 Marked

It displays research tasks marked as favorites, with special identifiers for easy recognition and access.

You can add tasks to or remove tasks from the favorites list at any time, allowing flexible management of favorite records.

You can sort by favorite time, task priority, or completion progress, and keyword search functionality is provided to quickly locate favorite tasks.

User Settings

In the bottom-left corner of the ScienceOne platform main page, you can view user information, and view system notifications. Click the username in the bottom-left corner to pop up the user settings menu, where you can view available coins balance and packages, perform benefit redemption, view user center, view usage instructions, submit feedback, contact us, change system language,set system themes, and log out. Click "More Information" to further understand our platform information, view update logs, and review rules and agreements.

7.1 Available Coins

After clicking the username, your current available coins are displayed in the pop-up menu. Click "Available Coins" to enter the "Coin Center", where you can view coin overview and subscription information.

7.2 Benefit Redemption

Click the "Benefit Redemption" option to pop up the benefit redemption interface. Enter redemption codes to redeem gifts. Your redemption records can be viewed in the bottom-right corner.

7.3 User Center

Click the "User Center" option to enter the personal user center page.

The "Profile" includes two sections: "Account Information" and "Personal Homepage".

7.3.1 Account Information

You can view and manage your basic information here.

Information Display: The page clearly displays your profile including avatar, username, work, account ID, account, phone number, email address, and registration time.

Information Editing: For the username field, you can click the "Edit" button on the right to modify; for phone number and email fields, you can click the "bind" button on the right to bind corresponding contact methods; for account password fields, you can click the "reset" button on the right to update your account password.

Account Cancellation: The bottom of the page provides an "Delete Account" button for permanently deleting your account.

7.3.2 Personal Homepage

You can create your exclusive academic card here.

Information Display: The page displays your basic information, educational background, and research/work experience.

Information Editing: In the basic information section, you can click the "Edit" button to modify your name and add/remove research areas; in the education section, you can click "Add Education" in the bottom-right to add educational experiences; in the research/work experience section, you can click "Add Research&Work Experience" in the bottom-right to add experiences, with support for "Delete" and "Edit" for each experience.

7.4 Channel Management

Connect communication platforms to get Feed subscriptions.

Click your username, select Channel Management in the pop-up menu, then choose the channel you want to connect.

7.4.1 WeChat

Click the Connect WeChat button, then scan the QR code with WeChat to finish connecting.

7.4.2 Feishu

• Log in to the Developer Console - Feishu Open Platform with your Feishu account. After successful login, click [Create Custom App].

• Fill in the app name and description, upload an app icon, then click [Create] to enter the app management page.

• On the left sidebar of the app management page, locate and click [Add Features]. Select [Bot] and then click Add.

• Click Credentials & Basic Info on the left sidebar. Locate the two parameters App ID and App Secret on the page, click the Copy button next to each value, and save them to your notes or memo.Please store these credentials securely and avoid leakage for data security purposes; they will be required in subsequent steps.

Click [Create Version] at the top to publish the application.

Open Feishu and enter the chat window of the newly created app. Click the Settings button, find the Chat ID parameter on the page, click the adjacent Copy button, and save the value to your notes or memo.Please store this ID securely and avoid leakage for data security purposes; it will be required in subsequent steps.

• Access the ScienceOne , click your username and select Channel Management in the pop-up menu. Click the Connect Feishu button, enter the Feishu Bot App ID, App Secret and Chat ID obtained in Step 4 above, then click Confirm Creation.

• On the left sidebar of the Feishu app management page, click Events & Callbacks. On the Event Configuration tab, select Receive Events via Long Connection and click Save.

• Click Add Event on the Event Configuration page. Search for and add the [Message received] event in the pop-up list, click Add, then follow the on-screen prompts to authorize required permissions.

• Enable recommendation permissions.
• On the Callback Configuration page under Events & Callbacks, select [Receive callbacks through persistent connection] as the subscription mode and click Save. No additional address input is needed; the configuration will take effect automatically.

• Click Permission & Scopes on the left sidebar, then click [Batch import/export scopes] on the page to open the permission import window.

• Copy the code block below, replace the original JSON content in the pop-up window, click [Next, Review New Scopes], then click Apply for Authorization. Wait for the permission import process to complete after confirmation.

{
  "scopes": {
    "tenant": [
      "im:message",
      "im:message.p2p_msg:readonly",
      "im:message.group_at_msg:readonly",
      "im:message:send_as_bot",
      "im:resource",
      "contact:user.base:readonly",
      "im:message.group_msg",
      "im:message:readonly",
      "im:message:update",
      "im:message:recall",
      "im:message.reactions:read",
      "docx:document:readonly",
      "drive:drive:readonly",
      "wiki:wiki:readonly",
      "bitable:app:readonly",
      "task:task:read",
      "contact:contact.base:readonly",
      "docx:document",
      "docx:document.block:convert",
      "drive:drive",
      "wiki:wiki",
      "bitable:app",
      "task:task:write"
    ],
    "user": []
  }
}

• On the left sidebar of the Feishu app management page, click [Version Management & Release], then click Create Version in the upper right corner.

• Enter an app version number (1.0.0 is used as an example here; you may customize your own version number) and release notes, then click [Save] and Confirm Release.

• Wait for Feishu administrators to approve the release. Once published successfully, you can start private chats with the Feishu Bot.

7.5 Tutorials

Select the "tutorials" option to view the system's detailed user tutorials, understanding system functions and usage methods, with support for switching between Chinese and English bilingual viewing.

7.6 Send Feedback

Select the "Send Feedback" option to submit opinions and suggestions about the system, supporting online feedback within the platform.

7.7 Language Settings

ScienceOne supports switching between Chinese and English system languages. In the user settings menu, click Language, and select your desired language from the list. The interface will automatically switch to that language version.

7.8 Theme Settings

After clicking the username, click "Theme" in the pop-up menu, then select theme options such as "Quantum White", "Dark Matter", or "System Theme" to switch the system's overall visual style. You can choose different themes based on personal preferences for a more comfortable user experience.

7.9 Contact Us

Select the "Contact Us" option to scan the code and join the official community for inquiries and assistance.

7.10 More Information

7.10.1 About US

Select the "About Us" option to view the system introduction, includingFeatures, WorkflowUse cases, For institutions, Foundation.

7.10.2 Release Notes

Select the "Release Notes" option to view system version update logs, with support for viewing main content updates for each version.

7.10.3 Rules & Agreements

Select the "Rules & Agreements" option to view various rules and agreements, including ScienceOne's Terms of Services, Privacy Policy, Institutional Account Activation & Verification Agreement, and Coin Rules.

7.11 Logout

Click the "Logout" button at the bottom of the menu to securely exit your current account. The system returns to the login page. For account security, it is recommended to log out after completing operations.

Notification System

On the system's main page, click the bell button in the bottom-left corner to pop up the notification center menu. In this menu, you can view and handle messages, follow up on task reminders, and receive system notifications.

8.1 Notification Center

The notification center displays "All" messages by default. You can click the "Unread" tab to quickly filter all notifications that have not been viewed. If there are no unread messages, this view displays "No unread messages".

Each unread message has a red dot identifier on its right side. After you click the message and complete corresponding interaction (such as jumping to a page or viewing a popup), the red dot automatically disappears, indicating the message has been read.

At the top of the notification center, you can find the "Read All" button. After clicking, all unread messages are immediately marked with read dot, will (including those on entry icons) disappear.

8.2 Instant Popups

To ensure you can promptly understand the latest progress of key tasks, the system sends reminders through instant popups in the upper-right corner of the interface. These reminders proactively inform you of task success, failure, or states requiring your confirmation, and guide you for subsequent operations.

Popup Composition and Interaction:

Each popup consists of a status icon, message content, and two buttons in the bottom-right: "View" and "Read".

• Status Icons: Divided into success, failure, and confirmation categories, helping you quickly identify message nature.
• View: Clicking directly jumps to the task interface associated with the message.
• Read: Clicking marks the message as read, closes the popup, and updates red dot reminders in the message center accordingly.

Popup Types and Duration:

Based on message importance, popups are divided into two types:

• Notification Messages (Success/Failure):

These popups are for result notifications, remaining on screen for several seconds before automatically disappearing to minimize disruption to your current operations.

• Confirmation Messages:

These popups involve key nodes requiring your decision, remaining persistently displayed on the interface until you actively click buttons to process them.

Coin System

9.1 Coin Center

Click the username in the bottom-left corner of the ScienceOne platform, then click "User Center" in the pop-up menu to enter the "Coin Center".

The "Coin Center" includes "Coin Overview", "Activity Center" and "Plans & Top-Ups" modules.

"Coin" serves as the virtual currency system for experiencing the Panshi website, usable only within this site. Based on different validity durations, it is divided into three types:

• Daily Coin: Valid on the day obtained, expires at 24:00 on acquisition day.
• Time-Limited Coin: 30-day validity period, expires at 24:00 thirty days after acquisition.
• Permanent Coin: Permanently valid during normal account usage.

9.2 Coin Overview

"Coin Overview" allows you to obtain overall coin-related information on the ScienceOne, including your membership information, current available research value, and historical research value details.

9.2.1 User

Personal information displays your membership information. You can understand all benefits under your current membership level here.

9.2.2 Available Coins

This displays your current available coins, including daily coins, time-limited coins, and permanent coins.

9.2.3 Coin Details

Coin details display all your historical records involving coin acquisition and consumption. You can choose to view detailed coin consumption or acquisition, and also support searching to view details within some time periods.

9.3 Activity Center

ScienceOne provides a series of activities to acquire coins, including daily check-ins, inviting friends, newbie tasks, etc. You can participate in various activities to acquire coins and use various ScienceOne functions.

9.3.1 Daily Login Activity

All users receive coins upon daily login.

9.3.2 Inviting Friends

Following guidelines to invite friends to register on ScienceOne, both can receive certain research value rewards.

9.3.3 New User Tasks

Complete new user tasks to receive time-limited coin rewards.Click the new user task card in the bottom-left corner of the ScienceOne platform to view the task list.

Click any task, follow the operation guide to complete the new user task and receive corresponding time-limited coin rewards.

9.4 Plans & Top-Ups

This Plans & Top-Ups module opens recharge for different memberships and displays benefit comparisons. You can flexibly choose different membership based on actual needs.

9.4.1 Basic Membership

ScienceOne provides Basic Membership services for all users. The first month after registration comes with 200 free coins, plus 2GB literature library space and 200 reference papers.

9.4.2 Plus Membership

The Plus Membership provides richer quota packages based on the Basic Membership.

9.4.3 Pro Membership

The Pro Membership not only has the richest quota packages but also provides priority access to new features.

9.4.4 Enterprise Plan

Enterprise users can contact business for exclusive customized research solutions, including dedicated deployment, quota customization, and exclusive custom domain services.