Multi-Model Support
ThinkEx integrates multiple AI providers, giving you flexibility to choose the best model for your needs.Available Models
- Gemini (Google)
- Claude (Anthropic)
- ChatGPT (OpenAI)
Gemini Models
- Gemini 3.0 Pro Preview: Latest preview model with advanced capabilities
- Gemini 3.0 Flash Preview: Latest fast preview model
- Gemini 2.5 Pro: Powerful and reliable for complex tasks
- Gemini 2.5 Flash (Default): Fast and efficient for everyday use
- Extended context windows
- Thinking mode for complex reasoning
- Native file processing
- Web grounding capabilities
Switching Models
Change the AI model at any time:The default model is Gemini 2.5 Flash, which provides a good balance of speed and capability for most tasks.
Workspace Context Awareness
The AI assistant has full awareness of your workspace content, making it truly context-aware.What the AI Knows
- All cards: Notes, PDFs, flashcards, videos, and more
- Content: Text, transcripts, extracted PDF text
- Structure: Folders, card relationships, layout
- Metadata: Card names, types, creation dates
Context Injection
ThinkEx automatically provides workspace context to the AI:- Workspace metadata: Name, description, and structure
- Selected cards: Cards you explicitly select for context
- Text selections: Highlighted text from notes or PDFs
- Reply context: Previous messages you’re responding to
AI Tools and Capabilities
The AI assistant has access to powerful tools for working with your workspace.Content Creation Tools
Create Notes
Generate new notes from prompts or sources
Create Flashcards
Build study decks with Q&A cards
Create Quizzes
Generate interactive quizzes with explanations
Add Videos
Search and add YouTube videos
Content Processing Tools
Process URLs
Extract content from web pages
Process Files
Upload and analyze documents
Web Search
Search the internet for information
Read Workspace
Access and summarize workspace content
Organization Tools
Update Cards
Edit and update existing cards
Search Workspace
Find content across all cards
Rename Items
Change card names
Create Folders
Organize cards into folders
Chat Features
Thread Management
Each workspace can have multiple chat threads:- Automatic threads: New threads created for different conversations
- Thread history: Access previous conversations
- Thread titles: Automatically generated based on content
- Archive threads: Remove old conversations from view
Message Types
The AI can generate different types of responses:- Text: Standard conversational responses
- Tool calls: Actions performed on your workspace
- Thinking: Reasoning process (for supported models)
- Citations: Source references with links
- Code: Syntax-highlighted code blocks
- Math: Rendered LaTeX equations
Context Features
Selected Cards Context
Selected Cards Context
Select cards before asking questions to focus the AI’s attention:
- Hold Cmd/Ctrl and click cards to select
- Selected cards are highlighted
- Ask your question
- The AI receives full content of selected cards
Text Selection
Text Selection
Highlight text in notes or the chat to reference it:
- Select text in a note card or AI message
- Your selection is automatically included in context
- Ask questions about the selected text
Reply Context
Reply Context
AI Gateway
ThinkEx uses an AI Gateway to route requests to different model providers.Gateway Benefits
- Unified interface: Single API for multiple providers
- Automatic routing: Requests sent to appropriate provider
- Prefix-based: Model IDs like
google/gemini-2.5-flash - Fallback support: Automatic retries if a provider fails
Model ID Format
Model IDs follow the pattern:provider/model-name
google/gemini-2.5-flashanthropic/claude-sonnet-4.5openai/gpt-5-chat
For backward compatibility, unprefixed Gemini models (like
gemini-2.5-flash) are automatically prefixed with google/.Advanced Features
Thinking Mode
Supported Gemini models can show their reasoning process:- Included thoughts: See how the AI approaches problems
- Thinking levels: Minimal, moderate, or extensive (model-dependent)
- Transparent reasoning: Understand the AI’s decision-making
Token Optimization
ThinkEx optimizes token usage for cost-efficiency:- Message pruning: Removes old reasoning and tool calls
- Context window: Manages conversation length
- Selective context: Only includes relevant workspace content
File Attachments
Attach files directly in chat:- Drag and drop files into the chat
- Supported formats: images, PDFs, documents
- Files processed with native AI capabilities
Common AI Commands
Here are examples of what you can ask the AI:Content Creation
Content Processing
Organization
Analysis
Best Practices
Effective Prompting
Effective Prompting
- Be specific about what you want
- Provide context by selecting relevant cards
- Break complex tasks into smaller steps
- Use clear, descriptive language
Model Selection
Model Selection
- Use Gemini Flash for quick, everyday tasks
- Use Gemini Pro for complex reasoning
- Try Claude for long-form writing
- Experiment with different models for different tasks
Context Management
Context Management
- Select relevant cards before asking questions
- Keep conversations focused on one topic
- Start new threads for different subjects
- Archive old threads to reduce clutter
Workspace Integration
Workspace Integration
- Use descriptive card names for easier AI reference
- Organize content in folders by topic
- Keep related content in the same workspace
- Let the AI help with organization tasks