Signpost AI Logo
App sectionsKnowledge

Knowledge Sources

Overview

Knowledge Sources are the foundation of your knowledge system - they represent the actual data that your AI agents can search and reference through collections. Sources can be files uploaded from your computer or live data connections to external systems.

1. Sources Dashboard

Knowledge Sources Interface Knowledge Sources Interface - Main dashboard showing all your data sources

The sources dashboard (SourcesTable component) provides:

  • Sources Table: Paginated table showing source name, type, last updated date, tags, and vector status
  • Inline Editing: Click edit button to modify source names directly in the table
  • Search Functionality: Search sources by name
  • Source Actions: Upload Files and Add Live Data buttons in the header
  • Row Actions: View and Delete options in dropdown menu for each source

2. File Upload Sources

File Upload Process

Knowledge Sources File Upload Knowledge Sources File Upload - Upload documents directly from your computer

  1. Click "Upload Files" from the sources dashboard
  2. File Upload Modal: Opens the FilesModal with tabbed interface
  3. Choose Upload Method:
    • File Upload Tab: Individual file selection and upload
    • Folder Crawler Tab: Upload multiple files from folder structure

File Upload Tab (FileUploadTab)

File Selection:

  • File Input: Select multiple files using browser file picker
  • File Parsing: Uses useFileParser hook for content extraction
  • Progress Tracking: Shows upload progress percentage during processing

File Management:

  • File List: Preview selected files with names and content
  • Name Editing: Customize source names before upload (defaults to filename)
  • Custom Tags: Add additional tags beyond default "File Upload" tag
  • Tag Management: Add/remove custom tags with tag input field

Upload Configuration:

  • Automatic Tags: Files automatically get "File Upload" tag plus file extension
  • Team Association: Sources automatically associated with current team
  • Database Storage: Content stored in sources table with metadata

3. Live Data Sources

Live Data Integration Process

Knowledge Sources Live Data Knowledge Sources Live Data - Configure real-time data feeds and API integrations

  1. Click "Add Live Data" from the sources dashboard
  2. Live Data Modal: Opens the LiveDataModal with form interface
  3. Select Integration Type: Choose from available live data source types

Available Live Data Types

The system supports several live data integration types:

Support Desk Integration:

  • Subdomain: Your support desk subdomain
  • Email: Support desk account email
  • API Token: Support desk API authentication token
  • Locale: Language setting (default: en-us)

Web Scraping (WebScrapingForm):

  • URL: Website URL to scrape
  • Processing Options: Text extraction settings
  • Include URLs: Option to include URL metadata
  • Extract Media Content: Process media elements
  • Retrieve Links: Extract link information

CMS Integration:

  • URL: CMS instance URL
  • API Token: CMS API authentication
  • Map Configuration: Field mapping settings

Prompt Integration (PromptForm):

  • Prompt: Custom prompt for data generation
  • Processing Configuration: Token limits and chunk settings

Bot Logs Integration (BotLogsForm):

  • Bot Log Configuration: Integration with bot logging systems
  • Data Extraction: Automated log content processing

4. Source Management Features

Sources Table Interface

Table Columns:

  • Name: Source name with inline editing capability (click edit icon)
  • Type: Source type (File, Support Desk, Web Scraping, etc.)
  • Last Updated: Date when source was last modified
  • Tags: Visual tags showing source category and metadata
  • Vector: Indicator showing if source has vector embeddings (Yes/No)

Inline Name Editing

Edit Process:

  1. Start Editing: Click the edit icon next to any source name
  2. Edit Mode: Input field appears with current name
  3. Save/Cancel: Use checkmark to save or X to cancel
  4. Keyboard Shortcuts: Enter to save, Escape to cancel
  5. Database Update: Changes saved to sources table immediately

Tag System

Automatic Tags:

  • File Upload: Added to all uploaded files
  • Live Data: Added to all live data integrations
  • File Extensions: File type extensions added automatically
  • Custom Tags: User-defined tags during upload

Tag Display:

  • Color Coding: Blue for "File Upload", Purple for "Live Data"
  • Visual Design: Rounded pill-style tags with appropriate colors
  • Tag Management: Tags managed through upload/creation process

5. Source Preview and Actions

Source Preview Interface (SourcePreview)

Preview Features:

  • Source Details: View complete source information and content
  • Content Display: Full text content of the source
  • Metadata Viewing: Creation date, last updated, tags, and type information
  • Source Actions: Edit, update, or manage source settings

Access Method:

  • Table Click: Click on any source name in the table to open preview
  • Row Selection: Preview opens in modal or side panel
  • Source Management: Update source information from preview interface

Source Deletion

Delete Process:

  1. Actions Menu: Click the three-dot menu for any source
  2. Select Delete: Choose delete option from dropdown
  3. Confirmation Dialog: DeleteSourceDialog confirms deletion intent
  4. Database Cleanup: Source removed from sources table
  5. Collection Impact: Source automatically removed from any collections

Delete Confirmation:

  • Warning Message: Clear indication of permanent deletion
  • Confirmation Required: User must explicitly confirm action
  • Cascade Effects: Automatic removal from collections and relationships

Database Structure

Sources Table Schema:

  • Primary Fields: id, name, type, content, team_id
  • Metadata Fields: tags, created_at, last_updated, vector
  • Team Isolation: All sources filtered by current team
  • Vector Storage: Embeddings stored as JSON in vector field

6. Processing Options and Configuration

Common Processing Options (ProcessingOptions)

Text Processing Settings:

  • Chunk Size: Configure how content is split (default varies by source type)
  • Chunk Overlap: Set overlap between text chunks for better context
  • Max Token Limit: Maximum tokens per chunk for API compatibility
  • Processing Quality: Options for content extraction quality

Live Data Configuration (CommonFields)

Standard Configuration Fields:

  • Name: Descriptive name for the data source
  • Type: Source type (automatically set based on integration)
  • Enabled: Toggle to enable/disable data source
  • URL: Primary endpoint or URL for data retrieval

Processing Configuration:

  • Advanced Options: Chunk size, overlap, and token limit settings
  • Content Extraction: Options for media content and link extraction
  • Update Frequency: How often to check for new data

7. Error Handling and User Feedback

Upload Process Feedback

File Upload Status:

  • Progress Tracking: Real-time progress bar during file upload
  • Success Indicators: Confirmation when files are successfully added
  • Error Messages: Clear error reporting for failed uploads
  • Validation Feedback: Immediate feedback for missing required fields

Live Data Integration Feedback

Configuration Validation:

  • Field Validation: Real-time validation of required fields
  • Connection Testing: Immediate feedback on API connectivity
  • Configuration Errors: Clear error messages for invalid settings
  • Success Confirmation: Confirmation when live data source is created

Database Integration

Team Isolation:

  • Automatic Team Assignment: Sources automatically associated with current team
  • Access Control: Users can only see sources from their team
  • Data Security: Team-based filtering ensures data privacy

React Query Integration:

  • Cache Management: Efficient data loading and cache invalidation
  • Real-time Updates: Automatic refresh when sources are modified
  • Performance Optimization: Paginated loading for large source lists