TypeAgent Knowledge Settings

AI Model Configuration

Checking AI model availability...

Agent Server Host

WebSocket URL of the TypeAgent agent server. Leave blank to use the default (ws://localhost:8999/). The browser agent's port is discovered automatically.

Default Extraction Mode

Choose the default mode for knowledge extraction operations

Basic Mode

Fast metadata extraction without AI - perfect for bulk operations

URL Analysis Domain Classification Basic Topics

Content Mode

AI-powered content analysis with entity and topic extraction

AI Content Analysis Entity Extraction Topic Identification

Actions Mode

AI analysis plus interaction detection for dynamic pages

AI Content Analysis Action Detection Interactive Elements

Full Mode

Complete AI analysis with relationships and cross-references

Full AI Analysis Relationship Extraction Cross-references

Batch Processing

Max Concurrent Extractions

1 (Slower) 3 10 (Faster)

Quality Threshold

0.1 (Permissive) 0.3 1.0 (Strict)

Action Auto-Discovery

Automatically discover available actions on page navigation

Discovery Mode

Skip sensitive sites (banking, healthcare, login pages)

Chat Panel

Open chat panel automatically for TypeAgent actions

When off, context-menu commands such as "Show actions on this page" still dispatch to TypeAgent, but the side panel will not pop open. The explicit "Open TypeAgent Chat" menu item is unaffected.