What this section is for
Knowledge Management Systems (KMS) allow creating, organizing, and maintaining knowledge bases for AI agents. Agents use this knowledge to provide accurate and up-to-date answers to users based on uploaded documents and articles.
How to access this section
- Through sidebar menu: "Knowledge" in navigation panel
- From home page: "My knowledge bases" section → "View All"
- Direct link:
/kms - Required permissions: Authorized user with knowledge base creation rights
Knowledge base list (/kms)
Page structure
Breadcrumbs and actions
- Navigation: Home → Knowledge management
- Main action: "Create knowledge base" button for creating new base
Knowledge base display
List format: Cards with main information
- Icon: Database (database icon)
- Name of knowledge base
- Description (if specified)
- "Open" button — go to base management
Loading states
- Skeleton screens — placeholders during data loading
- Adaptive quantity of cards based on screen height
- Hover effects for interactivity
Creating knowledge base (/kms/create)
Creation form
Required fields
Display Name
- Knowledge base name for display
- Used in interface and lists
- Required field for creation
Additional fields
Description
- Brief description of knowledge base purpose
- Helps in organization and search
- Displayed in base list
Automatic settings
Default values are set during creation:
- Ownership: Company — organization ownership
- Visibility: Corporate — visibility within organization
- Status: Published — published base
Creation process
- Fill required fields
- Click "Create KMS"
- Automatic redirect to created knowledge base
- Ready to add content
Knowledge base management structure (/kms/:id)
Section navigation
After creating knowledge base, tab system is available:
Articles
- Main content of knowledge base
- Creating and editing articles
- Publication status management
Files
- Uploading documents of various formats
- Automatic processing and indexing
- Linking with created articles
Settings
- Knowledge base configuration
- Text processing parameters
- Visibility settings
Article management (/kms/:id/articles)
Control panel
Search and creation functions
- Search field — by title and content (in development)
- "New Article" button — create new article
- Adaptive layout of control elements
Article list
Display: Table view with main information
- Article title — main identifier
- Status — published/draft/deleted
- Creation/modification date
- Actions — edit, delete
Pagination
- 10 articles per page by default
- Page navigation with large number of articles
- Adaptive control elements
Article deletion
Soft Delete with undo option
- Click delete button on article
- Notification appears with "Undo" button
- 5 seconds to cancel action
- Automatic final deletion after timer
Article restoration
- "Undo" button in notification
- Immediate article restoration
- Preserve all data and metadata
File management (/kms/:id/files)
Drag & Drop interface
Supported formats
- Text documents: PDF, DOC, DOCX, TXT
- Presentations: PPT, PPTX
- Spreadsheets: XLS, XLSX
- Other formats depending on configuration
Upload process
- Drag files to upload area
- Or click to select files from system
- Automatic upload to server
- Content processing and indexing
- Create linked articles from file content
Uploaded file management
- File list with size and type information
- Delete button for each file
- Go to article created from file
- File processing statuses
Upload limitations
- Maximum size: 20MB by default
- Number of files — limited by pricing plan
- File types — according to system configuration
Knowledge base settings (/kms/:id/settings)
Main parameters
Base information
Visibility
- Private — only for creator
- Corporate — for entire organization
- Public — open knowledge base
Name
- Updateable field for changing display name
- Required value
Description
- Additional information about base purpose
- Helps organize multiple bases
Text processing settings
Split Options
Enable splitting:
- Toggle switch for activation
- Function description and impact on search quality
Split Type:
- Token — split by AI model tokens
- Character — split by character count
Token split settings
Chunk Size:
- Number of tokens in one fragment
- Range: from minimum to maximum value
- Step: 100 tokens
- Default: according to CHUNKING_DEFAULTS
Chunk Overlap:
- Number of tokens for overlap between fragments
- Improves connectivity during search
- Range: 0-1000 tokens
- Step: 10 tokens
Character split settings
Separator:
- Separator characters for text splitting
- Input field with length limit
- Tooltip with usage hints
- Separator examples: newline, period, paragraph
Validation and limitations
- Minimum and maximum values for all parameters
- Automatic correctness checking of settings
- Warnings for incorrect values
Common usage scenarios
Scenario 1: Creating support knowledge base
- Go to creation: /kms → "Create knowledge base"
- Fill form:
- Name: "Support Knowledge Base"
- Description: "Frequently asked questions and instructions"
- Create base → go to management
- Upload documents: Files → upload FAQ, instructions
- Create articles: Articles → write additional materials
- Configure processing: Settings → enable token splitting
Scenario 2: Import company documentation
- Create new base for documentation
- Mass file upload: Files → drag & drop multiple documents
- Wait for processing — system will create articles from files
- Edit articles — improve structure and formatting
- Configure search parameters — optimize chunk size
Scenario 3: Managing existing base
- Open base from list
- View articles — Articles → analyze content
- Add new materials — Files or Articles → new content
- Delete outdated information — with undo option
- Update settings — Settings → adjust parameters
Scenario 4: Organizing content by topics
- Create thematic articles in Articles section
- Structure information by categories
- Use descriptive titles for better search
- Link files and articles through common theme
Important limitations
- File size: maximum 20MB per file
- Number of knowledge bases limited by pricing plan
- Supported formats — only popular office formats
- File processing may take time depending on size
- Article search is in development
- Knowledge base deletion — irreversible operation
Common problems and solutions
Problem: File doesn't upload
Solutions:
- Check file size (maximum 20MB)
- Ensure file format is supported
- Check stable internet connection
- Try refreshing page and uploading again
Problem: Article wasn't created from file
Solutions:
- Wait for file processing completion (may take several minutes)
- Check that file contains readable text
- Ensure file isn't corrupted
- Try uploading different format file
Problem: Can't find article in list
Solutions:
- Check pagination — article might be on another page
- Ensure article wasn't accidentally deleted
- Search function not available yet — check all pages
- Refresh page to update data
Problem: Text splitting settings don't save
Solutions:
- Ensure values are within allowed ranges
- Check that all required fields are filled
- Wait for save completion (loading indicator)
- Refresh page and check saved values
Problem: Knowledge base doesn't appear in agent list
Solutions:
- Check base visibility settings (should be Corporate or Public)
- Ensure base contains articles or files
- Wait for indexing update (may take time)
- Contact administrator to check access rights
Performance optimization
Content structure recommendations
- Split large documents into thematic articles
- Use descriptive titles to improve search
- Avoid information duplication between articles
- Regularly update outdated content
Text processing settings
- For short documents: increase chunk overlap
- For technical documentation: use character splitting
- For FAQ: optimal chunk size 500-800 tokens
- Experiment with settings for better search quality
File organization
- Group files by topics before uploading
- Prepare files: remove excessive formatting
- Upload in stages — not all files simultaneously
- Monitor processing of each file
What's next?
After creating and configuring knowledge base:
- Connect to agents — configure base usage in AI agents
- Create articles manually — detailed content management
- Configure tags — organize and categorize materials
- Monitor usage — analyze knowledge base effectiveness