Skip to main content

Notebooks - AI Chat Sessions

Notebooks are the core of your Bike4Mind experience. Each notebook is a persistent conversation with AI that you can return to, search, and build upon over time.

Getting Started

Creating a New Notebook

  1. Click New in the sidebar or press the new notebook button
  2. Start typing your message in the input area
  3. Press Enter to send (or Shift+Enter for a new line)

Your notebook is automatically saved and will appear in your sidebar for easy access.

  • Sidebar - All your notebooks appear in the left sidebar
  • Favorites - Star important notebooks for quick access
  • Search - Find notebooks by name or content
  • Projects - Organize related notebooks into projects

Input Options

Text Input

The rich text editor supports:

  • Markdown formatting - Bold, italic, lists, code blocks
  • Code highlighting - Automatic syntax detection
  • Math equations - LaTeX/KaTeX support
  • @mentions - Reference agents with @agent-name

File Attachments

Attach files to provide context for your conversation:

Upload Methods:

  • Click the attachment button
  • Drag and drop files into the chat
  • Paste from clipboard
  • Connect Google Drive for cloud files

Supported File Types:

  • Documents: PDF, DOCX, TXT, MD, HTML
  • Data: CSV, JSON, XLSX
  • Images: PNG, JPG, GIF, WebP
  • Code files: Most programming languages

Image Input

Upload images for AI analysis:

  • Single or multiple images per message
  • Image preview with lightbox gallery
  • AI can describe, analyze, and answer questions about images
  • Image editing tools available for supported models

Voice Input

Start a voice session for hands-free interaction:

  • Real-time speech-to-text transcription
  • Multiple voice options for AI responses
  • Visual audio feedback while speaking
  • Mute/unmute controls

Working with Messages

Message Actions

Every message has a menu with useful actions:

ActionDescription
CopyCopy the full response to clipboard
EditModify and resend your prompt
PinMark important messages for quick reference
DeleteRemove message from conversation
ForkCreate a new notebook branching from this point
DownloadSave as Markdown, text, or PDF

Code Blocks

AI responses often include code with:

  • Syntax highlighting for 100+ languages
  • One-click copy button
  • Expand for full-screen viewing

Rich Content

Responses can include:

  • Diagrams - Mermaid charts rendered visually
  • React components - Interactive UI elements
  • Charts - Data visualizations with Recharts
  • HTML/CSS - Styled content previews

AI Settings

Customize how AI responds in each notebook.

Model Selection

Choose from multiple AI models across providers:

  • OpenAI - GPT-5 family, GPT-4.1, O-series reasoning models
  • Anthropic - Claude 4.6 Opus, Claude 4.5 Sonnet/Haiku, and more
  • Google - Gemini 3, Gemini 2.5 Pro/Flash
  • xAI - Grok 4, Grok 3 family
  • Open models - Llama 4, DeepSeek, and more via AWS Bedrock

See AI Models for the full model guide. Filter models by type (Text, Image, Video), speed, or price tier.

Temperature

Control response creativity:

  • 0.0 - Focused, deterministic responses
  • 0.7 - Balanced (recommended for most uses)
  • 1.0+ - More creative, varied responses

Context & Memory

  • History lines - How much conversation context to include
  • Mementos - AI remembers important details across sessions
  • Knowledge files - Attach documents for reference

Special Modes

  • Quest Master - Break complex tasks into subtasks
  • Research Mode - Compare multiple models simultaneously
  • Artifacts - Enable rich interactive outputs

Advanced Features

Search Within Notebook

Use the search bar in the header to:

  • Find specific messages by text
  • Filter to show only pinned messages
  • Navigate long conversation histories

Layout Options

Customize your workspace:

LayoutBest For
Chat onlyFocused conversations
Horizontal splitChat + knowledge viewer
Vertical splitSide-by-side comparison
Picture-in-pictureMultitasking

Agents

Attach specialized AI agents to your notebook:

  • Each agent has a unique personality and expertise
  • Mention agents with @agent-name in your messages
  • Agents can proactively contribute to conversations

Smart Tools

Your notebooks have access to a full suite of AI-powered tools that extend what the AI can do beyond text generation. Toggle between Smart mode (AI auto-selects the right tools) and Fast mode (no tools, fastest response).

Available tool categories include:

  • Web Search & Deep Research - Find current information and conduct multi-step research
  • Knowledge Base Search - Reference your uploaded documents
  • Image Generation & Editing - Create and modify images
  • Charts & Diagrams - Generate data visualizations and Mermaid diagrams
  • Math & Calculations - Evaluate complex expressions
  • Astronomy - Moon phases, sunrise/sunset, ISS tracking, planet visibility
  • Blog Publishing - Draft, edit, and publish blog posts
  • External Integrations - Atlassian (Jira/Confluence), GitHub, and other MCP tools

See Smart Tools for the full guide on every available tool.


Keyboard Shortcuts

ShortcutAction
EnterSend message
Shift+EnterNew line in message
?Open Help Center
@Mention an agent

Tips & Best Practices

Writing Effective Prompts

  1. Be specific - Clear instructions get better results
  2. Provide context - Attach relevant files or explain background
  3. Iterate - Use edit and retry to refine responses
  4. Use agents - Specialized agents for specific tasks

Organizing Your Work

  1. Use Projects - Group related notebooks together
  2. Pin important messages - Quick reference later
  3. Star favorites - Access frequent notebooks quickly
  4. Download key responses - Save important outputs

Managing Long Conversations

  1. Fork when starting a new direction
  2. Search to find specific exchanges
  3. Pin milestones and key decisions
  4. Create snippets to preserve important context

Troubleshooting

Message Not Sending

  • Check your internet connection
  • Verify you have available credits
  • Try refreshing the page

Slow Responses

  • Consider using a faster model
  • Reduce the amount of context/history
  • Check if complex tools are enabled

File Upload Issues

  • Verify file size is under limits
  • Check file type is supported
  • Try a different upload method