AI Chat Guide

Overview

This chapter explains the shared AI chat capability in Unitary Studio Web and helps you understand:

• which pages expose AI chat,
• how AI chat covers both everyday conversation and execution-oriented operations,
• how AI chat can expand its handling scope through built-in capabilities such as skills when supported,
• how to choose among Ask, Plan, and Agent modes,
• how to send text and image attachments and stop an in-progress answer when needed,
• how to configure a custom model service.

This chapter is useful when:

• you want to interact in natural language before manually navigating through product pages,
• you are already on a feature page and need help understanding parameters, task states, or recommended next steps,
• you want to attach screenshots so AI can analyze both the text and the image content,
• you want AI to help with controlled operations such as task creation or file handling.

Entry Points and Usage

AI chat is no longer attached to only one page. The home page and several feature pages expose an AI chat entry so you can ask questions or trigger supported actions with the current page context.

• It can be used for ordinary natural-language conversation.
• In execution-oriented scenarios, it can also assist with actual operations through controlled internal capabilities.
• On feature pages, current page context often makes the answer or available actions more relevant to what you are doing.

Workspace Selection

On pages that support quantum programming, a workspace selector is available at the top of the AI chat panel. Use it to switch which quantum programming workspace the current conversation is associated with.

• Click the workspace selector to view all available workspaces under your account.
• After selecting a workspace, the AI chat context switches to that workspace. Subsequent code execution and file operations will all be based on the selected workspace.
• If you have not created any workspace yet, the selector provides a quick entry to create one.

The workspace selector only appears on quantum programming pages. It is not shown on the home page, PDE solver, or other non-programming pages.

Basic Interface

On pages that support AI chat, you will usually see a panel like the following. This figure focuses on the core controls that are visible by default:

Numbers in the figure correspond to:

1. Fullscreen button: click it to expand the current AI chat panel into a larger fullscreen editing area.
2. New chat button: click it to start a new conversation and clear the current context before asking again.
3. Input box: type a question and press Enter or click the send button to submit.
4. Image attachment button: click it to pick local images and send them together with the current question.
5. Skills button: opens the available skills or built-in capability entry so the current chat can use more specific handling abilities.
6. Chat mode selector: supports Ask, Plan, and Agent modes.
7. Workspace selector: switch the quantum programming workspace associated with the current conversation. Only visible on quantum programming pages.
8. Model selector: switch the AI model that answers.
9. Send button: submit the current input.

The figure below shows the AI chat area after one round of Q&A, with focus on the send / stop control in the bottom-right corner.

Numbers in the figure correspond to:

1. Stop button

Additional notes:

• After you click the input, the chat area expands automatically to show more context.
• After sending, the button switches to Stop, meaning the answer is still being generated.
• Once a round is finished, you can collapse the panel back to its default height.

History Management

After sending a message to AI on the main page, the system automatically displays all historical conversations under your account, helping you efficiently organize and review past interactions.

• View history: The chat panel automatically shows all historical conversations under the current account. Each entry is displayed with its conversation title for quick identification.
• Review conversations: Click any history entry to fully review all Q&A, work ideas, and collaboration results from that conversation, enabling seamless workflow continuity.
• Delete conversations: Delete individual conversations you no longer need. Click the action button next to an entry and choose delete to remove invalid or redundant content, keeping your conversation list clean and organized.

Suggestions:

• Regularly clean up completed or unnecessary conversations to keep the list easy to browse.
• Keep important conversation records after finishing work for later review and reference.

Chat Modes

AI chat supports the following three modes:

• Ask: best for direct Q&A, such as clarifying a parameter or asking what a page element means.
• Plan: best for step-by-step suggestions, such as planning how to complete a PDE workflow.
• Agent: best for execution-oriented or continuous assistance; when the scenario supports it, this mode can call internal APIs, skills, and related capabilities to perform actual actions such as creating tasks, querying tasks, handling files, or advancing multi-step flows.

Use them like this:

• Choose Ask when you mainly want an explanation.
• Choose Plan when you want a structured sequence of steps.
• Choose Agent when you want the system to help execute or continue handling the task.

Keep in mind:

• Agent mode does not simply mean “a more detailed answer”; it may trigger controlled system operations.
• Before allowing execution-oriented actions such as task creation or file handling, make sure the current page and target are correct.

Skills Capability

When the scenario supports it, AI chat can also invoke skills to extend how it handles a request. You can think of skills as prebuilt capability packages that are better suited to certain classes of tasks than plain free-form conversation alone.

• skills are useful for tasks that follow a repeatable pattern or need a more specialized handling path.
• Click the Skills button next to the input area to view the skills currently available in the chat.
• Typing / in the input box also opens the skills picker; the picker supports multi-select so you can combine multiple capabilities in one request.
• The skills you can actually see and choose from are the capabilities already configured by the system, so the picker content is the current usable scope for that conversation.

If the response starts behaving more like structured task handling, tool usage, or guided execution, that usually means AI is already using one of the underlying internal capabilities instead of returning only plain text.

Weekly Paper Scenario Q&A

AI chat has been deeply optimized for the Weekly Paper scenario. The intelligent Agent can precisely link to the current paper page content, enabling a “what you see is what you ask” experience.

How to use:

• Open the AI chat panel while on a Weekly Paper page, and the Agent automatically senses the document content on the current page.
• Ask questions directly about the paper’s content, data, and key points — no need to manually copy and paste text or import files.
• AI interprets, analyzes, and expands on the paper’s content using the full context of the current page.

Typical scenarios:

• Daily literature reading: ask questions about paper content and quickly understand core ideas and methods.
• Weekly report review: ask AI for summaries, analysis, and suggestions based on page data.
• Document analysis: perform deep follow-up questions and extensions on tables, data, and conclusions.

The page-linking capability for Weekly Paper depends on the content structure of the current page. If the page content changes significantly, refresh the page before starting new questions.

Sending Image Attachments

AI chat supports sending images together with the current question.

Numbers in the figure correspond to:

1. Image attachment button: click it to pick local images and send them together with the current question.
2. Attachment preview area: shows the images currently attached so you can enlarge, review, or remove them before sending.
3. Question input: combine text instructions with attached images so AI can use both textual and visual context.

Suggestions:

• Keep the image focused on the problem you are asking about.
• If you provide both text and images, AI uses both in the same request.
• You can also paste screenshots directly from the clipboard into the input box.
• If too many images are added at once, only the first 10 images are kept and a notice is shown.

Custom Model Settings

If you want AI chat to use your own model service, open Model Settings from the avatar menu in the top-right corner.

After the modal opens, refer to the image below for each configuration item.

Numbered callouts in the image:

1. Enable model switch: located at the bottom-left of the modal; when enabled, the current custom model configuration can be used, and when disabled, the system default service is used.
2. Service URL field: enter the http or https endpoint of your custom model service.
3. API Key field: enter the credential required by the custom model service.
4. Model name field: enter the model identifier used in actual requests.
5. Save Configuration button: after validation passes, click it to save; the configured model is appended to the model list, and selecting it uses the custom model service.

Notes:

• When Enable model is turned on, the service URL, API Key, and model name must all be filled in, and the service URL only accepts http or https.
• When the switch is not enabled, the saved configuration is kept for later reuse.
• The configuration is stored in the current browser environment. If you switch browsers or clear local data, you may need to enter it again.

Download Confirmation

In Agent mode, when the system is about to download a generated result or a workspace file to your local device, it first shows a download confirmation card asking whether to download the file. The download is never executed automatically.

Numbered callouts in the image:

1. Download confirmation card: shows the file name, path, and confirmation timeout, and asks whether to download the file.
2. Approve Download button: confirms and starts downloading the current file.
3. Reject Download button: rejects the current confirmation and does not download the file.

Additional notes:

• If you do nothing before the countdown ends, the request expires and AI must initiate it again.
• If the request was approved but the actual file fetch fails, the card keeps the approved state and provides a Retry Download action.
• The download confirmation applies on a per-download request basis. Whether a file is downloaded is always under your control, and the system does not execute it automatically.

Context Auto-Compression

To increase the depth of interaction within a single conversation and break through traditional turn limits, the system includes an intelligent context auto-compression mechanism.

How it works:

• During a conversation, the system automatically monitors context length and intelligently trims redundant information when approaching the limit.
• The compression process preserves core content without losing meaningful information, while significantly freeing up conversation capacity.

What this means for you:

• You can ask more follow-up questions in a single conversation, with each question building coherently on the last.
• Complete complex problem decomposition, deep exploration, and multi-dimensional expansion in one session without frequently starting new conversations.
• This mechanism is transparent to you — no manual intervention is needed. The system handles compression and optimization automatically in the background.

Context auto-compression is a built-in system behavior and does not affect the core logic or answer quality of the current conversation. If you notice that answers in a long conversation no longer align with earlier context, compression may have been triggered — starting a new conversation at that point is the better choice.

Tool Permission Management (Agent Mode)

In Agent mode, the system performs fine-grained permission management for callable tools, distinguishing between read-only tools and operational tools.

Tool categories:

• Read-only tools: used only for content querying, viewing, and reading. They do not involve any modification, editing, or state-changing operations. Examples include viewing file listings, reading task status, and querying algorithm information.
• Operational tools: involve creation, modification, deletion, execution, or other state-changing operations. Examples include creating tasks, modifying configurations, and deleting files.

Approval rules:

• Read-only tools are fully exempt from the approval process. The Agent can invoke them directly without requiring confirmation each time.
• Operational tools maintain the approval confirmation mechanism. Before executing a modifying operation, the Agent displays a confirmation card for you to decide whether to proceed.

What this means for you:

• Routine query and browsing operations are instantly available with zero friction — no need to click confirm repeatedly.
• The approval steps are greatly streamlined, reducing unnecessary interruptions and improving team productivity.
• Data security and permission standards remain intact — all operations involving modification or deletion still require your explicit authorization.

The system automatically determines whether a tool is read-only based on its actual capabilities. You do not need to configure anything manually. If you are unsure about a tool’s permission classification, watch for the confirmation card: no confirmation card means the tool is read-only.

When AI Chat Fits Best

• You want to communicate in natural language instead of manually finding a specific button or form first.
• You do not understand a parameter, task status, or result and want a plain-language explanation first.
• You want to attach a screenshot so AI can analyze the problem region together with the current page context.
• You want step-by-step guidance or want the system to perform a controlled action such as creating a task, querying a task, handling a file, or invoking a matching skill.

Related Chapters

• Want the shortest first-visit path: see Quick Start.
• Want to understand how AI chat is used specifically inside PDE workflows: see PDE Solver Guide.
• Want to enter the code space or the quantum-programming workflow: see Quantum Programming Guide and Code Space Guide.