AI Chat

Interact with powerful AI models to analyze documents, answer questions, and boost productivity.

Three Ways to Chat:
1. AI Lab - Direct chat with any AI model
2. Knowledge Base Chat - AI that knows your documents
3. Greatest Minds - Chat with AI personalities

AI Lab - Multi-Model Chat

The AI Lab (Ctrl+1) provides direct access to multiple AI models for general-purpose conversations.

enTANGlement - AI Lab
šŸ¤–
AI Lab
šŸ“š
Docs
šŸ”
Search
āš™ļø
Settings

AI Lab

Model:Claude 3 Sonnetā–¼
šŸ¤–

Select a model above to start chatting

Available Models

Mistral 7B

Fast, efficient

LocalFree

Llama 3.2

Balanced performance

LocalFree

Claude 3 Sonnet

Advanced reasoning

ProCloud

GPT-4o

Creative & versatile

ProCloud

Grok

Real-time knowledge

ProCloud

Using AI Lab

  1. Select a Model - Click the model selector at the top of the chat
  2. Start Chatting - Type your message and press Enter
  3. Switch Models Mid-Conversation - Change models anytime to compare responses
  4. Save Important Chats - Pin conversations for quick access
enTANGlement - AI Lab Chat

AI Chat

Model:Claude 3 Opusā–¼
U

Explain quantum computing in simple terms

2:34 PM
AI
Mistral 7B

Quantum computing is like having a magical coin that can be both heads and tails at the same time, until you look at it. This lets quantum computers try many solutions simultaneously...

2:35 PM
U

Can you provide more technical details?

2:36 PM
AI
Claude 3 Sonnet

Certainly! Quantum computers use quantum bits (qubits) that leverage superposition and entanglement. A qubit can exist in a state |ĻˆāŸ© = Ī±|0āŸ© + Ī²|1āŸ© where Ī± and Ī² are complex probability amplitudes...

2:37 PM
Ctrl+Enter to sendā€¢Shift+Enter for new line

Knowledge Base Chat

When you're in a knowledge base, the AI has full context of your documents.

Key Features

šŸ“š
Document-Aware
AI searches your documents to provide accurate, sourced answers
šŸ”
Citation Links
Click citations to jump to source documents
šŸ’­
Context Preservation
Maintains conversation context across sessions
šŸŽÆ
Focused Responses
Answers based on your specific documents

Example Use Cases

enTANGlement - Knowledge Base Chat

AI Chat

Model:Claude 3 Opusā–¼
U

What are the key findings from the Q3 reports?

3:42 PM
AI
Claude 3 Sonnet

Based on the Q3 reports in your knowledge base: ā€¢ Revenue increased 23% YoY to $4.2M [Q3-Financial.pdf, p.2] ā€¢ Customer acquisition cost decreased by 15% [Marketing-Report.pdf, p.8] ā€¢ New product launch exceeded targets by 40% [Product-Update.pdf, p.1]

3:43 PM
Ctrl+Enter to sendā€¢Shift+Enter for new line

Document Context

climate_report.pdf95%

Page 42: "Global temperatures have risen..."

analysis.docx78%

Section 3.2: "Environmental factors include..."

research_notes.md72%

Line 145: "Key findings suggest that..."

Knowledge Base
šŸŒ

Climate Research

42 documents ā€¢ 2.3 GB

Pro Tip: Use specific document names in your questions for more focused answers: "What does Report2024.pdf say about sustainability?"

Greatest Minds - AI Personalities

Chat with AI personalities modeled after historical figures, experts, and fictional characters.

šŸ§ 
Einstein

Physics & relativity

šŸŽØ
Da Vinci

Art & invention

šŸ“š
Shakespeare

Literature & poetry

šŸ”¬
Marie Curie

Chemistry & physics

šŸ’”
Tesla

Electricity & innovation

šŸ›ļø
Socrates

Philosophy & logic

šŸŽÆ
Sun Tzu

Strategy & tactics

šŸš€
Carl Sagan

Astronomy & science

How Personalities Work

  1. Each personality has unique knowledge and communication style
  2. Responses reflect the historical figure's expertise and era
  3. Can be combined with your knowledge bases for specialized analysis
  4. Perfect for creative brainstorming and unique perspectives

Chat Features

Chat History

  • Automatic Saving - All conversations are saved automatically
  • Search History - Find past conversations by content
  • Pin Important Chats - Keep frequently referenced chats accessible
  • Archive Old Chats - Clean up without deleting
  • Export Conversations - Save as Markdown or PDF

Streaming Responses

Responses stream in real-time, allowing you to:

  • See partial answers immediately
  • Stop generation if the response goes off-track
  • Save time on long responses

Keyboard Shortcuts

New chatCtrl+Shift+N
Search historyCtrl+H
Stop generationEsc
Copy last responseCtrl+Shift+C
Switch modelCtrl+M
Toggle sidebarCtrl+B

Advanced Features

Multi-Model Conversations

Compare responses from different models side-by-side:

Model Comparison
āš”

Mistral 7B

Response time: 0.8s

Machine learning is a way for computers to learn patterns from data without being explicitly programmed for each task.

šŸ§ 

Claude 3 Sonnet

Response time: 1.2s

Machine learning is a subset of artificial intelligence that enables computers to automatically learn and improve from experience. It uses algorithms to find patterns in data and make predictions or decisions.

  1. Ask a question to one model
  2. Switch models and ask the same question
  3. Compare responses to find the best answer
  4. Combine insights from multiple models

Context Management

Control how much context the AI remembers:

  • Full Context - AI remembers entire conversation (slower)
  • Recent Context - Last 10 messages (balanced)
  • Minimal Context - Last 2 messages (fastest)

Custom Instructions

Set persistent instructions for AI behavior:

  • Response format (bullet points, paragraphs, code)
  • Language preferences
  • Technical level (beginner, intermediate, expert)
  • Tone (formal, casual, academic)

Privacy & Security

Local AI (Ollama)

100% Private: When using local models, your conversations never leave your computer. No internet connection required after model download.

Cloud AI (Pro)

When using cloud providers:

  • Only your messages are sent to the provider
  • Document content stays local (only relevant excerpts sent)
  • Conversations are encrypted in transit
  • No data is stored by enTANGlement servers
  • You can delete provider data anytime

Important: Cloud providers (OpenAI, Anthropic, xAI) have their own privacy policies. Review them before sharing sensitive information.

Optimizing Chat Performance

For Speed

  • Use smaller, local models (Mistral 7B)
  • Reduce context window size
  • Clear chat history periodically
  • Close unused knowledge bases

For Quality

  • Use larger models (Claude, GPT-4)
  • Provide clear, specific prompts
  • Include relevant context in questions
  • Use custom instructions for consistent output

Troubleshooting

Common Issues

IssueSolution
AI not respondingCheck model is downloaded/API key is valid
Slow responsesSwitch to faster model, reduce context
Generic answers in KB chatEnsure documents are processed, be more specific
Chat history not savingCheck storage permissions, restart app
Model switching failsVerify model is installed/configured

Tips & Best Practices

Effective Prompting

  • Be specific - "Summarize chapter 3" vs "Summarize the document"
  • Provide context - "As a Python developer, how do I..."
  • Ask follow-ups - Build on previous responses
  • Request formats - "List as bullet points" or "Explain with examples"

Managing Conversations

  • Start new chats for different topics
  • Pin reference conversations
  • Export important insights
  • Archive completed projects

Related Topics