Chat Widget

The AIO Chat Widget (aio_chat_widget) module provides a generic, reusable AI chat window component integrated natively into the OpenEduCat interface. It enables users to interact with AI agents directly from any page in Odoo via a top navigation bar (systray) icon, featuring real-time message streaming, markdown rendering, and context-aware assistance.

Key Features

  • Systray Quick Access: Dedicated AI icon in the top navigation bar for immediate access from any screen.

  • Flexible Window Management: Docked floating window with controls to minimize, expand to fullscreen, or close.

  • Real-time Message Streaming: Progressive text rendering as the AI generates responses, minimizing wait time.

  • Rich Markdown & Code Rendering: Supports formatted text, bullet lists, tables, and code snippets with XSS security.

  • Session & Context Awareness: Automatically retains chat history and active model/record context during the session.

Interface & Workflow

1. Launching from Systray

  • Click the AI Chat Icon located in the top-right navigation bar (systray) of the Odoo interface.

  • The floating chat window opens in the bottom-right corner of your screen.

../../_images/cw1.png ../../_images/cw2.png

2. Window Controls & Management

Use the header icons on the chat window to control the workspace:

../../_images/cw3.png

  • Minimize: Collapse the window into a compact dock at the bottom screen edge.

../../_images/cw6.png

  • Fullscreen: Expand the chat interface for a full, distraction-free view.

../../_images/cw4.png ../../_images/cw5.png

  • Clear Chat: Reset the current conversation context and start a fresh session.

  • Close: Hide the chat window.

3. Real-Time Message Streaming & Markdown

  • Live Streaming: Responses stream chunk-by-chunk in real time with auto-scrolling to keep up with incoming text.

  • Rich Formatting: Bold text, lists, tables, and fenced code blocks render cleanly with syntax formatting.

../../_images/cw7.png

Backend Architecture & Integration

  • Execution Endpoint: Communicates directly with the backend aio.agent service using synchronous playground execution (execute_agent_sync).

  • Context Payload: Passes model name (res_model) and active ID (res_id) to provide context-aware AI assistance for the current record.

  • Error Resilience: Catches network or execution errors gracefully and displays user-friendly toast notifications and in-chat error logs.