Updated on 2026-07-30 GMT+08:00

R&D Office

The R&D Office function focuses on non-coding scenarios in the daily routine of R&D personnel. Leveraging the intelligent chat mode of CodeArts Agent Space, you can easily complete tasks such as document writing, requirement sorting, business analysis, work reporting, and team collaboration, significantly improving office efficiency.

Five Major Application Scenarios

  • Intelligent technical document generation: Based on code context and prompts, the system automatically organizes logic and generates clearly structured technical documents, eliminating tedious manual formatting.
  • Deep insight into R&D data: The system automatically collects and analyzes R&D process data to generate visual reports, providing a quantitative basis for team efficiency assessment and project progress monitoring.
  • Rapid Product Requirement Document (PRD) construction: The system converts vague business ideas into rigorously structured PRDs through chat interaction, accelerating requirement implementation.
  • One-click report generation: The system integrates project progress, key milestones, and data achievements to quickly generate structured reporting materials, making weekly and monthly reports effortless.
  • Seamless mobile collaboration: The system deeply integrates with instant messaging (IM) to support initiating collaboration, approvals, and communication anytime and anywhere from mobile devices, ensuring real-time information synchronization and efficient decision-making in remote scenarios.

Intelligent Technical Document Generation

CodeArts Agent Space features built-in intelligent generation capabilities for design documents and technical proposals, allowing for the rapid output of well-structured and comprehensive documentation.

  1. Access CodeArts Agent Space by referring to Enabling CodeArts Agent Space.
  2. Enter a task prompt. Click Doc Generation. Type your task into the box, and click or press Enter to submit your request.

    If you need to generate technical documents based on existing materials, first click and select Upload Attachments to upload files, and then enter a task prompt.
    Write a guide on office backend permission allocation for new employees. It must cover steps, precautions, and FAQs, ensuring that the language is easy to understand and the format is standard.

    In addition to manual task input, CodeArts Agent Space also provides you with pre-configured templates for frequently used prompts. You can click to automatically fill them into the input box to help you get started quickly.

    You are a rigorous senior backend engineer, expert in transforming technical solutions into actionable, detailed design documents.
    You are a veteran technical architect, expert in drafting clear, well-structured technical proposals. Based on the requirement description, generate a comprehensive technical solution.
    You are a backend engineer specialized in API design and fluent in RESTful specifications. Based on the described features, generate a RESTful-compliant API document.

  3. View the generated results. After the task is complete, a document in Markdown format will be automatically generated under Editor > EXPLORER.

    Figure 1 View technical documents

  4. Confirm the results. View the information of the generated files and operate according to your actual needs, or modify them directly based on the generated content.

    In CodeArts Agent Space, context changes are presented through a Diff view, making the modified content more intuitive and easy to understand. The chat interface will display the total number of files, lines of code, and the file list involved in this change. Clicking View Changes after any file allows you to view the specific modifications in the Diff view, helping you clearly grasp every detail of the changes.

    Choose either of the following methods to confirm the files as required:

    • Method 1: quick operations on the chat interface

      Hover the mouse over the row where the file is located, and click or on the right to accept or reject the file changes, respectively. If there are multiple files, you can click Accept All to accept all files or Reject All to reject all files.

    • Method 2: operations in the explorer

      Under Editor > EXPLORER, view the content of the generated file. Click Keep (Ctrl+Y) on a segment to accept the generated content, or click Undo (Ctrl+N) to reject the generated content. Alternatively, at the bottom of the file, click Keep All (Ctrl+Enter) or Undo All (Ctrl+Delete) to accept or reject the generated file, respectively.

      If the editor is not displayed properly, click (Toggle Tool Panel) in the upper right corner of the top menu to open the tool panel. On the displayed panel, click Editor.

In-depth Insight into R&D Data

CodeArts Agent Space allows you to upload only spreadsheets (.xlsx, .xls, and .csv), documents (.ppt, .pdf, .doc, .docx, .md, and .txt), and images (.jpg, .png, .jpeg, and .webp) for data analysis, document generation, and more.

  • Images: The size of a single image cannot exceed 10 MB, and the total size of all images cannot exceed 20 MB.
  • Non-image attachments: A maximum of 20 attachments can be uploaded at a time, with no restrictions on the attachment size.
  1. Access CodeArts Agent Space by referring to Enabling CodeArts Agent Space.
  2. Enter a task prompt. Choose Data Analysis. Click , select Upload Attachments, and upload local files to be analyzed. Type your task into the box, and click or press Enter to submit your request.

    Perform an overall data analysis of this month's sales data, organize the changes in revenue, sales volume, and customer orders, summarize the overall business performance, and provide a brief conclusion.

    In addition to manual task input, CodeArts Agent Space also provides you with pre-configured templates for frequently used prompts. You can click to automatically fill them into the input box to help you get started quickly.

    You are a data engineer focused on data governance, specialized in cleaning dirty data. Write Python code to process the data based on user requirements.
    You are a data visualization expert proficient in Matplotlib and Seaborn, adept at recommending optimal chart types based on data characteristics. Write Python code to render charts as required.
    You are a senior data analyst skilled in exploratory data analysis using Python (Pandas). Write Python code to fully diagnose the uploaded data file and output a report.

  3. View the data analysis results. After the task is complete, a document in Markdown format will be automatically generated under Editor > EXPLORER.

    Figure 2 Viewing data analysis results

  4. Confirm the results. View the information of the generated files and operate according to your actual needs, or modify them directly based on the generated content.

    In CodeArts Agent Space, context changes are presented through a Diff view, making the modified content more intuitive and easy to understand. The chat interface will display the total number of files, lines of code, and the file list involved in this change. Clicking View Changes after any file allows you to view the specific modifications in the Diff view, helping you clearly grasp every detail of the changes.

    Choose either of the following methods to confirm the files as required:

    • Method 1: quick operations on the chat interface

      Hover the mouse over the row where the file is located, and click or on the right to accept or reject the file changes, respectively. If there are multiple files, you can click Accept All to accept all files or Reject All to reject all files.

    • Method 2: operations in the explorer

      Under Editor > EXPLORER, view the content of the generated file. Click Keep (Ctrl+Y) on a segment to accept the generated content, or click Undo (Ctrl+N) to reject the generated content. Alternatively, at the bottom of the file, click Keep All (Ctrl+Enter) or Undo All (Ctrl+Delete) to accept or reject the generated file, respectively.

      If the editor is not displayed properly, click (Toggle Tool Panel) in the upper right corner of the top menu to open the tool panel. On the displayed panel, click Editor.

Rapid PRD Construction

The system converts vague business ideas into rigorously structured PRDs through chat interaction, accelerating requirement implementation.

  1. Access CodeArts Agent Space by referring to Enabling CodeArts Agent Space.
  2. Enter a task prompt. Click PRD Generation. Type your task into the box, and click or press Enter to submit your request.

    If you need to generate requirement documents based on existing materials, first click and select Upload Attachments to upload files, and then enter a task prompt.
    Write a complete PRD for the intra-city errand ordering feature of a community lifestyle app, covering product objectives, user roles, business workflows, UI prototype specifications, field definitions, interaction rules, exception scenarios, and priority levels.

    In addition to manual task input, CodeArts Agent Space also provides you with pre-configured templates for frequently used prompts. You can click to automatically fill them into the input box to help you get started quickly.

    You are a veteran product manager expert in clarifying complex needs and distilling business goals into well-structured PRDs. Generate a complete PRD draft from the requirement description, business background, or discussion content provided.
    You are a meticulous senior product manager skilled in refining raw requirements into PRDs. Using the provided requirement information or PRD draft as input, produce a complete, logically structured PRD ready for review, collaboration, and execution.
    You are a product manager focused on requirements analysis and functional design, adept at breaking vague ideas down into clear features, business rules, and acceptance criteria. Based on the function requirements, generate a structured specification document.

  3. Refine product requirements. To obtain more accurate results, select the desired effects in the list to further refine your requirements. After you submit your choices, the AI will generate clear and well-organized content accordingly.

    Figure 3 Refining product requirements

  4. View the generated results. After the task is complete, a document in Markdown format will be automatically generated under Editor > EXPLORER.
  5. Confirm the results. View the information of the generated files and operate according to your actual needs, or modify them directly based on the generated content.

    In CodeArts Agent Space, context changes are presented through a Diff view, making the modified content more intuitive and easy to understand. The chat interface will display the total number of files, lines of code, and the file list involved in this change. Clicking View Changes after any file allows you to view the specific modifications in the Diff view, helping you clearly grasp every detail of the changes.

    Choose either of the following methods to confirm the files as required:

    • Method 1: quick operations on the chat interface

      Hover the mouse over the row where the file is located, and click or on the right to accept or reject the file changes, respectively. If there are multiple files, you can click Accept All to accept all files or Reject All to reject all files.

    • Method 2: operations in the explorer

      Under Editor > EXPLORER, view the content of the generated file. Click Keep (Ctrl+Y) on a segment to accept the generated content, or click Undo (Ctrl+N) to reject the generated content. Alternatively, at the bottom of the file, click Keep All (Ctrl+Enter) or Undo All (Ctrl+Delete) to accept or reject the generated file, respectively.

      If the editor is not displayed properly, click (Toggle Tool Panel) in the upper right corner of the top menu to open the tool panel. On the displayed panel, click Editor.

One-Click Report Generation

The system integrates project progress, key milestones, and data achievements to quickly generate structured reporting materials, making weekly and monthly reports effortless.

  1. Access CodeArts Agent Space by referring to Enabling CodeArts Agent Space.
  2. Enter a task prompt. Click PPT Generation. Type your task into the box, and click or press Enter to submit your request.

    If you need to generate PowerPoint files based on existing materials, first click and select Upload Attachments to upload files, and then enter a task prompt.
    Generate a product introduction PowerPoint for CodeArts Agent, showcasing functional highlights, core advantages, application scenarios, and parameter specifications, with a minimalist tech style.

    In addition to manual task input, CodeArts Agent Space also provides you with pre-configured templates for frequently used prompts. You can click to automatically fill them into the input box to help you get started quickly.

    You are a senior consultant expert in distilling core insights into structured, logical presentation decks. Generate a complete PPT outline and slide-by-slide content based on the topic and materials provided.
    You are a rigorous presentation expert. Using the provided documents, data insights, or discussion notes, produce structured, high-impact slide copy ready for presentation.
    You are a content strategist specialized in business presentations, tailored for diverse audiences. Generate presentation content based on the topic, target audience, and purpose provided.

  3. View the generated results. After the task is complete, a PowerPoint file is automatically generated under Editor > EXPLORER. The file cannot be viewed online. You need to open it locally.

    Figure 4 Generating a PowerPoint file

    CodeArts Agent Space does not support online viewing of PowerPoint files. Go to the directory of the corresponding project and use local software to view the PowerPoint files.

Instant Connection on Mobile Devices

Leveraging the powerful connectivity of IM, the barriers to data and interaction between the PC and mobile devices are broken down. It aims to solve the pain point of fragmentation between focused creation and instant communication in remote working, achieving a seamless transition across the entire process from immersive creation on desktop to mobile instant collaboration. For details, see IM Integration.

Managing R&D Artifacts on the Artifact Panel

During task execution or after completion, you can view real-time information such as to-do items, deliverables, skills, model context protocol (MCP), and context utilization through the artifact panel.

Table 1 Management operations on the Artifact Panel

Operation

Description

Viewing to-dos

Click in the upper right corner of the chat window. You can view the to-do list of the current chat under To-Dos.

Viewing task artifacts

Click in the upper right corner of the chat window. You can view the final output of the current chat under Task Product. Click the generated file to view its content in Documents or Editor.

Once the task execution is completed, you need to first accept and confirm the output files. Then the files are displayed here.

Viewing context information

Click in the upper right corner of the chat window. You can view details about the skills and MCP tools invoked by the current chat under Context Information.

Viewing context usage

Click in the upper right corner of the chat window. You can view the context usage (percentage of used tokens) of the current chat under Context usage in real time. This helps you intuitively manage your current context usage and prevent disruptions to the session continuity due to exceeding limits.