Open Chat & Use General Mode
Assignee: All team members π¬
Loyal Professor is an AI assistant embedded directly into every page of AIPD β not a separate page, but a side panel that can be opened and closed at any time. Loyal Professor has two modes: General (free-form chatbot) and Assisting (querying the project's Knowledge Base).
This page explains how to open the chat and use General mode. For Assisting mode (Q&A over project documents), see Q&A Project Knowledge.
Step 1: Open the chatβ
There are two ways to open Loyal Professor:
| Method | Action |
|---|---|
| Floating button | Click the round "R" button in the bottom-right corner of the screen |
| Keyboard shortcut | Ctrl + Shift + L (Windows/Linux) or Cmd + Shift + L (Mac) |
The side panel slides in from the right (default width 550px). You can drag the left edge to resize, or click the handle to close it.
Step 2: Get familiar with the interfaceβ
The chat panel is divided into 3 sections, top to bottom:
Header β topβ
| Element | Purpose |
|---|---|
| Loyal Professor logo (left) | Click to open the mode dropdown (General / Assisting) |
| Conversation title (center) | Shows the current conversation's name. Click to rename inline β after typing, click β to save |
| Pencil icon | New Chat β start a new conversation |
| History icon | History β open a popover to view and search old conversations |
| Expand icon | Open workspace β open the chat in full-screen mode (if the page supports it) |
| X icon | Close β close the panel |
Message List β middleβ
Displays the message history:
- Your messages: bubble on the right, user avatar, HH:mm timestamp.
- AI messages: bubble on the left, sparkles avatar, content streams in realtime β a blinking cursor shows while the AI is composing.
When there are no messages yet, an empty state is shown with the title "How can I help?" and 4 preset suggestions β click a suggestion to auto-fill the input.
Input β bottomβ
The message input has the placeholder "Ask anything... (Type @ to add context)". It auto-expands as you type multiple lines (up to ~4 lines).
The toolbar on the right side of the input:
- @ button β "Add context (@)" β attach context from the project (used in Assisting mode).
- Send button (paper plane icon) β send the message.
Below the input is the line "Content quality may vary check important info." β a reminder to double-check the answers yourself.
Step 3: Choose a modeβ
Click the Loyal Professor logo in the top-left of the header to open the dropdown:
| Mode | Icon | Description | Use when |
|---|---|---|---|
| General | π¬ MessageCircle | "Free-form chat, no KB" | Asking general knowledge (coding, best practice, architecture) not related to a specific project |
| Assisting | π§ BrainCircuit | "AI assistant with knowledge base" | Asking about documents, tickets, or specs of the project you are currently working on |
Switching modes resets the current conversation β if you want to keep the history, create a new conversation before switching mode.
Step 4: Use General modeβ
General mode works like a regular ChatGPT/Claude β it does not query project documents, and answers based on the model's general knowledge.
Empty-state suggestions for General mode:
- "What are best practices for REST API design?"
- "How do I structure a microservice?"
- "Explain dependency injection in simple terms"
- "What's the difference between unit and integration tests?"
- "How can I improve code readability?"
- "What are common security pitfalls in web apps?"
Steps:
- Type your question into the input.
- Press Enter to send, or Shift + Enter for a new line.
- The AI answers with realtime streaming β text appears progressively.
- Keep typing follow-ups to go deeper β the AI remembers the context of the current conversation.
Step 5: Manage conversationsβ
Click the History icon in the header to open the conversation-list popover.
Features:
- Search box β filter by conversation title.
- Grouped by time: Today β Yesterday β This Week β This Month β Older.
- Relative timestamps: "Just now", "5m ago", "2h ago", "3d ago" or "Mar 15".
- Click a conversation to reload all of its messages.
- Click "Start New Chat" (when the list is empty) to start a new one.
Rename a conversation:
- Click the conversation title in the header.
- An inline input appears; type the new name (up to 255 characters).
- Press Enter or click the β icon to save. Press Esc or X to cancel.
Protecting sensitive informationβ
Before sending each message, the system automatically scans for credentials (API key, token, password, etc.) and replaces them with [MASKED_xxx].
- During scanning, the status bar shows "Checking for sensitive data...".
- If any credentials were masked, your message appears with a "Protected" badge β hover it to see how many credentials were protected.
Sensitive information is masked before it leaves the browser β the AI only receives the masked version. No need to worry about pasting logs, configs, or code snippets that contain secrets into the chat.
General mode is suitable when you need general knowledge that is not project-specific β e.g., learning a new concept, reviewing a pattern, or asking about syntax. If the question concerns code, documents, or specific tickets of the project you are working on, switch to Assisting mode for accurate answers based on the Knowledge Base β see Q&A Project Knowledge.