Knowledge Bases
Organize your documents into intelligent, searchable collections powered by AI.
What is a Knowledge Base?
A Knowledge Base is a collection of related documents that enTANGlement processes and indexes for intelligent search and chat. Each knowledge base is isolated, ensuring your personal documents stay separate from work projects.
Core Concepts
🗂️ Domain Isolation
Each knowledge base is completely isolated from others. Documents, chat history, and AI context remain separate.
🤖 Custom AI Models
Choose different AI models for each knowledge base. Use fast models for quick searches, powerful models for analysis.
📁 Folder Watching
Link a folder to automatically import new documents. Perfect for ongoing projects or research.
🔍 RAG Processing
Documents are processed using Retrieval-Augmented Generation for accurate, context-aware search results.
Creating a Knowledge Base
Method 1: Quick Start with Templates
enTANGlement offers pre-configured templates for common use cases:
AP Biology
High school advanced biology
Calculus I
University mathematics
World History
Historical documents & notes
Computer Science
Programming & algorithms
Chemistry
Chemical formulas & labs
Literature
Books & literary analysis
Method 2: Custom Knowledge Base
- Open Creation Dialog
- Click "Create New" button
- Or press
Ctrl+N - Or use
Ctrl+Kthen select "Create New"
- Configure Basic Settings
Name: My Research Project Description: Collection of papers on climate change Category: Professional Icon: 🔬 (optional)
- Select Storage Location
Choose where documents will be stored on your computer. enTANGlement will:
- Create the folder if it doesn't exist
- Watch for new files automatically
- Organize documents by type (PDFs, docs, etc.)
- Choose AI Model
Select the AI model for this knowledge base:
Model Speed Quality Best For Mistral 7B Fast Good Quick searches, summaries Llama 3.2 Medium Better Detailed analysis Claude 3 (Pro) Fast Excellent Complex reasoning GPT-4 (Pro) Medium Excellent Creative tasks
Pro Tip: You can change the AI model later in the knowledge base settings. Start with a fast model and upgrade when needed.
Managing Documents
Adding Documents
There are three ways to add documents to a knowledge base:
- Drag and Drop - Drag files directly onto the document area
- Browse and Select - Click "Add Documents" to open file picker
- Automatic Import - Files added to the linked folder are imported automatically
Import Modes
When adding documents, choose how they're managed:
| Mode | Description | Use When |
|---|---|---|
| 📄 Copy | Creates a copy in the KB folder | You want enTANGlement to manage files |
| ↗️ Move | Moves files to the KB folder | Organizing scattered documents |
| 🔗 Link | References files in original location | Files must stay in current location |
Document Processing
When documents are added, enTANGlement:
- Extracts Text - Pulls text from PDFs, Word docs, etc.
- Chunks Content - Splits into semantic sections
- Generates Embeddings - Creates vector representations
- Indexes for Search - Makes content searchable
Processing Documents
Searching Your Knowledge Base
Natural Language Search
Unlike traditional keyword search, enTANGlement understands what you mean:
| Traditional Search | enTANGlement Search |
|---|---|
| "climate change effects" | "What are the main impacts of global warming on agriculture?" |
| "python list comprehension" | "How do I filter a list in Python efficiently?" |
| "quarterly revenue 2024" | "Show me our financial performance this year" |
Search Operators
For advanced users, special operators are available:
type:pdf- Search only PDFsdate:2024- Documents from 2024author:"John Doe"- Specific authorhas:images- Documents with images
AI Chat Integration
Each knowledge base includes an AI assistant that knows your documents:
- Click the chat icon or press
Cin a knowledge base - Ask questions about your documents
- AI provides answers with source citations
Knowledge Bases
Climate Research
42 documents • 2.3 GB
CS Notes
18 documents • 456 MB
Q4 Reports
8 documents • 124 MB
Example Conversations
You: "Summarize the main findings from the climate reports" AI: Based on the 3 climate reports in your knowledge base... You: "Compare the methodologies used in paper1.pdf and paper2.pdf" AI: Looking at both papers, here are the key methodological differences... You: "Create a study guide for tomorrow's biology exam" AI: Based on your biology notes and textbook chapters, here's a comprehensive study guide...
Knowledge Base Settings
Configure your knowledge base by clicking the settings icon:
General Settings
- Rename knowledge base
- Update description and icon
- Change storage location
- Export/Import settings
AI Configuration
- Switch AI models
- Adjust chunk size for processing
- Configure embedding models
- Set context window size
Document Settings
- File type filters
- Auto-import rules
- Duplicate handling
- Processing priority
Advanced Features
Multi-Model Strategy
Use different models for different tasks within the same knowledge base:
- Embedding Model - For document indexing (MiniLM, BGE, E5)
- Search Model - For query understanding (Mistral, Llama)
- Chat Model - For conversations (Claude, GPT-4)
Automation Rules
Set up automatic actions for your knowledge base:
- Auto-categorize documents by content
- Generate summaries for new documents
- Alert when specific topics appear
- Export weekly reports
Best Practices
Organizing Knowledge Bases
- One topic per KB - Keep subjects separate for better AI context
- Consistent naming - Use clear, descriptive names
- Regular maintenance - Remove outdated documents
- Backup important KBs - Export regularly
Optimizing Performance
- Chunk size - Smaller chunks for technical docs, larger for narratives
- Model selection - Match model power to task complexity
- Document quality - OCR PDFs for better text extraction
- Regular reindexing - Refresh embeddings periodically
Storage Tip: Each knowledge base requires about 10-20% additional storage for embeddings and indexes. A 1GB document collection needs approximately 1.2GB total space.
Troubleshooting
Common Issues
| Issue | Solution |
|---|---|
| Documents not appearing | Check file format is supported, verify folder permissions |
| Slow search results | Reduce chunk size, use faster embedding model |
| AI gives generic answers | Ensure documents are fully processed, check model selection |
| High memory usage | Limit concurrent processing, use smaller models |