ai-assistant/AI Assistant Interface Guide

3.2 AI Assistant Interface Guide

The Pop AI Assistant interface is carefully designed for long-form conversations and quick access during workflow building, document editing, knowledge-base queries, and daily tasks.
This chapter helps you quickly understand the interface structure, available functions, and interaction methods.


🖥️ Overall Interface Layout

The AI Assistant interface is divided into three major sections:

  1. Left: Session List (Sidebar)
  2. Center: Main Chat Window
  3. Right: Settings / Information / Multi‑Window Control (Right Toolbar)

Each section is described below.


📌 1. Left Panel: Session List (Sessions)

The left panel displays all your conversations—regular chats, role-based sessions, temporary sessions, and pinned sessions.

Features include:

  • Create New Session
    Each session has an independent context and can use different models, roles, or knowledge bases.

  • Session naming & renaming
    Allows better organization.

  • Session search
    Quickly filter sessions by keyword.

  • Archive / Delete sessions
    Keep the list clean and organized.

  • Role session filtering
    Sessions created with specific "AI Roles" will show their avatar.


💬 2. Center: Main Chat Window

This is the most frequently used area for free-form conversation, multi-turn interaction, file uploads, and result previews.

The chat window supports:

● High‑quality Markdown rendering

  • Syntax-highlighted code blocks
  • Tables
  • Math formulas (LaTeX)
  • Lists / blockquotes

● File Upload

Upload images, PDFs, and text files for AI to analyze.

● Multi-turn context

Conversations extend naturally, with automatic context optimization to reduce token usage.

● Knowledge source citations

When using knowledge-base Q&A, the assistant returns the document sources.

● Message action buttons

Each message supports:

  • Copy
  • Ask follow-up
  • Rewrite / reflect (optional)

🧩 3. Right Toolbar (Functional Sidebar)

The right panel contains settings and advanced tools related to the current session.

● Model Selection

Switch between OpenAI, DeepSeek, Ollama, local models, or enterprise APIs.

● Role Selection

Switch AI personas during the conversation.

● Knowledge Base Binding

Choose whether the session can access your knowledge base.

● Multi‑Window Layout

Consistent with Pop Workflow layouts—open multiple chat windows side‑by‑side.

Examples:

  • Left: GPT‑4
  • Right: DeepSeek
  • Bottom panels for document analysis or code explanation

● Settings Drawer

Includes:

  • Temperature
  • Max output length
  • Output style (concise / detailed)
  • Preferred translation language
  • Markdown & formula toggles

🪟 Multi‑Panel View

Pop supports one‑click 1, 2, 3, or 4‑panel chat layouts.

Mode Description
Single Standard chat
Dual Model comparison / role discussion
Triple Document analysis + Q&A + summary
Quad Expert‑panel discussion / parallel tasks

Each panel can independently:

  • Select a model
  • Select a role
  • Bind a knowledge base
  • Maintain its own context

This is one of Pop AI Assistant’s signature features.


🎛️ Input Area (Message Input)

The message input bar supports:

  • Multi-line editing
  • File insertion
  • Screenshot insertion
  • Slash commands (/summary, /translate)
  • Role hints
  • Send (Enter) or newline (Shift + Enter)

Future updates will include “prompt templates” and “suggested completions.”


🧭 Top Navigation Bar

The top navigation displays:

  • Current session name
  • Model in use
  • Active role
  • Conversation statistics (tokens, request count)
  • AI processing status (loading indicator)

🔚 Summary

The AI Assistant interface is designed to be:

  • Clear and easy to use
  • Modular and extensible
  • Deeply integrated into Pop’s core scenarios (documents, workflows, dashboards)