Skip to main content

Overview

The chat interface is where you interact with AI models in real-time. It provides a rich set of features for conducting conversations, managing context, and collaborating with team members.
Conversation Interface
When you open a conversation, you’ll see a clean, intuitive interface designed for natural AI interaction. At the top, you’ll find your Space name and current assistant—in this case, the “Greeting” assistant. The conversation flows naturally in the center, with your messages and the AI’s responses clearly separated by timestamps. Each message from the AI shows which model was used (you’ll see a reference to “Model Guide & Comparison” that you can click for details). Below each AI response, you’ll find a helpful set of actions: copy the response, add it to a document, create a task from it, bookmark it, give feedback with thumbs up or down, and access additional options. You can even continue the conversation in different ways using the forward arrow. The input area at the bottom is where the magic happens. Here you can type your questions naturally, and the helpful placeholder text reminds you of powerful features—use @ to reference files from your space, tap into web search for current information, or target a specific AI model for your query. Quick-access buttons let you attach files, add images, or even use voice input. When you’re ready, just hit the Send button to continue the conversation. On the right side, you’ll find quick actions for managing your conversation: close it, adjust settings, switch assistants, bookmark for later, or copy the conversation to another space.

Interface Layout

Chat Input Area with Suggestions
The conversation interface is thoughtfully designed with three main areas that work together to give you a powerful, intuitive chat experience.

The Input Area

The input area is your command center for interacting with AI. It’s located at the bottom of the screen and packed with intelligent features to make your conversations more productive. Smart Suggestions Dropdown
Command Center with AI
As you start typing, especially when using the @ symbol, a helpful dropdown appears above your input showing relevant suggestions. This intelligent menu adapts to what you’re typing and offers:
  • Web Search - Trigger real-time web searches to get current information for your prompt. Perfect when you need facts, news, or up-to-date data.
  • Image Generation - Quickly access image generation tools like DALL-E 3 to create visuals based on your descriptions.
  • Space Search - Upload data to your space, and Space Search will automatically find relevant information from your documents to enhance your prompt with the right context.
  • Quick Assistant Access - Browse and switch between assistants on the fly. See options like Ally (your support agent), Booster (SEO assistant), Buzz (social media manager), Finny (business manager), and more. Each shows a brief description so you know exactly what they’re designed to help with.
  • Start typing to see more relevant suggestions appear dynamically based on your input.
The Text Input Field
Text Input Area with Features
The main text area where you compose your messages supports rich features:
  • Type naturally—the field expands as you write longer messages
  • Use @ to reference files from your space, making it easy to give the AI context from your uploaded documents
  • Trigger web search capabilities to pull in fresh information
  • Target specific models directly from your input
  • The placeholder text guides you: “Type your question here. Use @ to reference files, use web search or to target a specific model.”
Quick Action Buttons
Command Center with AI
Three essential buttons sit below the input field for fast access:
  • Attachment button (+) - Upload files, documents, images, or data to include with your message
  • Image button (picture icon) - Add images directly to your conversation for visual context
  • Voice input button (microphone icon) - Speak instead of type—perfect for hands-free interaction or when you want to explain something quickly
Send Button The prominent “Send” button on the right submits your message. You can also press Enter to send (customizable in settings).

Main Chat Area

The central conversation view is where your dialogue unfolds naturally. Messages flow chronologically with clear visual separation between you and the AI. Each of your messages shows your profile picture and timestamp, while AI responses display the assistant’s avatar and which model generated the response. When the AI uses tools or references sources, you’ll see clear indicators showing what data was accessed. Responses appear in real-time as they’re generated, giving you immediate feedback. Below each AI message, you’ll find action buttons to copy the response, add it to a document, create a task, bookmark it for later, give feedback, or explore more options. The right side of the interface houses powerful panels that slide out when you need them. These give you complete control over your conversation’s behavior and help you build your personal AI workspace. Click any icon in the top-right to reveal its panel:
  • Assistant Panel (robot icon) - Your gateway to AI personalities and capabilities
  • Models Panel (microchip icon) - Fine-tune which AI models are available
  • Data Panel (folder icon) - Manage your space’s knowledge base
  • Collections Panel (bookmark icon) - Your saved responses and personal knowledge library
  • Share Panel (share icon) - Collaborate with team members
  • Comments Panel (comment icon) - Add discussions and notes
Each panel slides smoothly into view, letting you make adjustments without disrupting your conversation flow.

Starting a Conversation

1

Navigate to Space

Open your space from the Spaces menu
2

Click Input Area

Click the “Start a new conversation” text field
3

Enter Your Prompt

Type your message or question
4

Submit

Press Enter (or Cmd/Ctrl+Enter) or click Send

Message Composition

Text Input

  • Multi-line: Press Shift+Enter for new lines
  • Formatting: Supports Markdown formatting
  • Mentions: Tag team members with @username
  • Commands: Use slash commands for quick actions

File Attachments

Attach files to provide context:
  • Documents: PDF, DOCX, TXT, etc.
  • Images: JPG, PNG, GIF, etc.
  • Data Files: CSV, JSON, XML, etc.
  • Code Files: Various programming languages
To attach:
  1. Click attachment icon or drag-and-drop
  2. Select files from your computer
  3. Wait for upload completion
  4. Files appear as chips below input
  5. Submit message with attachments

Code Blocks

Format code properly:
# Use triple backticks with language
def hello_world():
    print("Hello, World!")

Message Features

Message Actions

Each message has action buttons:
  • Copy: Copy message content
  • Edit: Modify your own messages
  • Delete: Remove messages
  • Share: Share specific messages
  • React: Add emoji reactions
  • Comment: Add threaded comments
  • Save: Add to collections

Regenerate Responses

If you’re not satisfied with a response:
  1. Click the regenerate button
  2. The AI generates a new response
  3. Compare different variations
  4. Choose the best version

Edit and Resend

Modify your prompts:
  1. Click edit on your message
  2. Make changes
  3. Resubmit to get new response
  4. Previous responses remain in history

AI Response Features

Streaming Responses

Responses appear in real-time:
  • Progressive Display: See text as it’s generated
  • Stop Generation: Cancel mid-response
  • Smooth Experience: Natural typing effect

Tool Call Display

When AI uses tools:
  • Tool Name: What tool is being used
  • Parameters: What data is being queried
  • Results: Output from the tool
  • Success/Failure: Tool execution status

References & Sources

AI responses can include:
  • Source Citations: Links to referenced material
  • Data Sources: Which files were consulted
  • Web Links: External resources used
  • Confidence Indicators: Reliability signals

Context Management

Conversation Context

The AI maintains context throughout:
  • Previous Messages: Entire conversation history
  • Attached Files: Uploaded documents
  • Space Data: Relevant space information
  • User Preferences: Your settings

Context Window

Understanding token limits:
  • Each model has a maximum context size
  • Older messages may be truncated
  • Important context is prioritized
  • Use summaries for long conversations

Clearing Context

Start fresh when needed:
  • Begin a new conversation
  • Explicitly state “ignore previous context”
  • Use a different conversation thread

Models Panel

Click the microchip icon to open the Models panel: Header
  • Title: “Models”
  • Tip: “You have full control over which models are available for use in your space. Select or deselect models to tailor your options, ensuring that only your preferred models are ready for generating responses.”
Model List
  • Checkboxes - Select/deselect models
  • Provider Icons - Visual identification (AI21, Anthropic, Google, OpenAI, etc.)
  • Model Names - Full model identifiers with versions
  • Checked models - Currently selected, indicated with blue checkmark
  • Scrollable - Long list of available models
  • Close button (X) - Hide panel
Using the Models Panel:
  1. Click microchip icon to open
  2. Browse available models
  3. Check/uncheck models to customize selection
  4. Selected models become available in conversations
  5. Close panel when done

Assistant Panel

Click the robot icon to open the Assistant panel: Empty State (No Assistant Selected):
  • Icon display area with gear/sparkle icon
  • Message: “No Assistant selected”
  • Sub-message: “Choose or create an assistant to personalize your experience”
  • Search bar: “Search for an assistant”
  • ”+ Create Assistant” button - Quick assistant creation
Model Recommendations: When open, shows “Based on this conversation, we suggest trying:” with recommended models based on context, each with:
  • Provider icon
  • Model name
  • Info icon (i) - View model details
  • Click to switch to that model
Search dropdown - “or select a model…” with searchable model list:
  • Grouped by provider
  • Provider icons and names
  • Full model list
  • Search functionality

Data Panel

Click the folder icon to open the Data panel: Header:
  • Title: “Data”
  • Tip: “Utilize labels to categorize related files and effortlessly reference them in your threads using @-refs”
  • Search bar: “Search all files”
  • Refresh icon - Reload file list
  • ”+ Add” button - Upload new files
  • “Labels” button - Manage file labels
File Area:
  • Drop zone: “Drop file(s) here” for drag-and-drop uploads
  • File list (when files are present)
  • File type icons
  • File names
  • Organization by labels
Using Data Panel:
  1. Click + Add to upload files
  2. Drag and drop files into drop zone
  3. Search for specific files
  4. Use @-refs in messages to reference files
  5. Organize with labels

Collections Panel

Click the bookmark icon to open the Collections panel: Header:
  • Title: “Collections”
  • Tip: “Find all your favorite messages in one place. Keep important notes handy by adding them to collections for quick access, and remove them when they’re no longer needed.”
  • Search bar: “Search all collections”
  • ”+” button - Create new collection
Empty State:
  • Message: “No results.”
  • Instructions to start adding messages
Using Collections:
  1. Click bookmark icon on messages to save
  2. Organize saved messages into collections
  3. Search collections quickly
  4. Access important information easily
  5. Remove when no longer needed

Share Conversation

Click the share icon to open share dialog: Dialog Content:
  • Title: “Share Conversation”
  • Sub-title: “Invite collaborators to join this discussion.”
Members Section:
  • Label: “Members with access”
  • Dropdown: “Select an organization member or type an email to invite”
  • Add team members to conversation
  • Grant conversation access
Share Link Section:
  • Label: “Share this link:”
  • URL field: Full conversation URL (copyable)
  • Copy button - Copy link to clipboard
  • Note: “This link only works for users who already have access to this Space.”
Actions:
  • Cancel - Close without changes
  • Share - Confirm sharing

Model Switching

Mid-Conversation Switching

Change models without losing context:
  1. Open Models panel (microchip icon)
  2. Check/uncheck models or use assistant panel suggestions
  3. Select different model from recommendations
  4. Continue conversation
  5. Context is preserved

Model Comparison

Compare model responses:
  1. Ask same question with different models
  2. Use multiple conversations
  3. Check Assistant panel for suggestions
  4. Evaluate quality differences
  5. Choose best model for task

Keyboard Shortcuts

Speed up your workflow:
ShortcutAction
EnterSend message (configurable)
Cmd/Ctrl + EnterNew line or send (configurable)
Shift + EnterNew line
Cmd/Ctrl + KFocus input
Cmd/Ctrl + /Show shortcuts
EscClear input or close panels
Edit last message
Cmd/Ctrl + FSearch conversation

Conversation Settings

Configure conversation behavior:

General Settings

  • Auto-save: Automatically save conversations
  • Streaming: Enable/disable real-time responses
  • Timestamps: Show/hide message times
  • Notifications: Alert preferences

Input Settings

  • Enter Key Behavior: Send or new line
  • Auto-complete: Suggest completions
  • Spell Check: Enable/disable
  • Voice Input: Speech-to-text

Display Settings

  • Theme: Light or dark mode
  • Font Size: Adjust readability
  • Message Density: Compact or comfortable
  • Code Highlighting: Syntax theme

Accessibility Features

Screen Reader Support

  • Semantic HTML for assistive technologies
  • ARIA labels on interactive elements
  • Keyboard navigation throughout interface

Visual Adjustments

  • High Contrast: Improved visibility
  • Font Scaling: Adjustable text size
  • Color Themes: Various options
  • Focus Indicators: Clear keyboard focus

Keyboard Navigation

  • Tab through interactive elements
  • Arrow keys for history navigation
  • Escape to close panels
  • Standard shortcuts throughout

Mobile Experience

The interface adapts for mobile:
  • Responsive Layout: Optimized for small screens
  • Touch Gestures: Swipe to access panels
  • Simplified UI: Essential features prioritized
  • Voice Input: Easy on mobile

Best Practices

  • Write clear, specific prompts
  • Provide necessary context
  • Break complex tasks into steps
  • Ask follow-up questions
  • Iterate on responses
  • Attach relevant documents early
  • Reference previous messages
  • Start new conversations for new topics
  • Summarize long conversations
  • Use collections for important content
  • Share conversations with team members
  • Use comments for discussions
  • Tag members when input needed
  • Keep conversations organized
  • Document decisions
  • Use keyboard shortcuts
  • Master model switching
  • Leverage templates
  • Save frequently used prompts
  • Organize conversations well

Troubleshooting

Common Issues

Solutions:
  • Check internet connection
  • Verify you’re logged in
  • Try refreshing the page
  • Check message length limits
  • Ensure files are uploaded
Solutions:
  • Check model load status
  • Try a faster model
  • Reduce context length
  • Check network speed
  • Simplify your prompt
Solutions:
  • Check conversation history
  • Reattach necessary files
  • Summarize previous context
  • Start new conversation if needed
  • Use explicit references
Solutions:
  • Check file size limits
  • Verify file format
  • Try different file
  • Check storage quota
  • Contact support if persists

Next Steps

I