Dashboard

Total Resolved
-
Loading...
Pending Drafts
-
Loading...
Acceptance Rate
-
Loading...
KB Articles
-
Loading...

Recent Activity

Loading activity...

System Status

All Systems Operational
Ollama Service Active
ChromaDB Active
Zammad API Active
Queue Processor Active

Quick Actions

Processed Tickets

All processed draft responses

Showing 0 of 0 tickets
Loading tickets...

Knowledge Base Explorer

Browse and search FAQ documents and ticket summaries

Showing 0 of 0 entries
Loading knowledge base...

Analytics

Analytics and reporting view will be implemented here.

UI Components Showcase

Sample elements and controls

Buttons

Toggles & Switches

Form Elements

Tags & Badges

Primary Success Warning Danger Info Neutral

Server Process Management

Control the main AI Support server process

Checking...

Main server running on port 3000

Changes to settings below require a server restart to take effect.

Zammad Configuration

Connection settings for your Zammad ticketing system

The base URL for your Zammad API endpoint
API token with necessary permissions
Comma-separated list of user IDs to ignore (staff and AI users)

LLM Configuration (Ollama)

Large Language Model settings

Ollama API endpoint URL
Model name to use for generation
Maximum context length (tokens)
Creativity level (0.0 - 2.0)
Leave empty if authentication not required

Knowledge Base Configuration

Vector database and embedding settings

Number of results to retrieve from KB
Model for vector embeddings
⚠️ Changing the embedding model requires re-embedding your knowledge base using python knowledge-base-py/re_embed.py all

Reranking Configuration

Second-stage ranking for improved search accuracy

Use a cross-encoder model to rerank search results
Cross-encoder model for reranking
Number of results after reranking

Server & Webhook Configuration

Main server port and webhook settings

Port for the main webhook server (server.js)
Port for the dashboard UI (ui-server.js)
Port for the inspector server (inspector-server.js)
Base URL to reach the inspector server (for nginx/reverse proxy setups)
Base URL to reach the webhook server (for nginx/reverse proxy setups)

Process Configuration

Parallel processing and batch settings

Concurrent ticket resolution processes
Concurrent ticket summarization processes
Number of tickets to fetch per batch

Draft Tracking Configuration

Settings for draft comparison and tracking

Similarity ≥ this value = accepted as-is (0.0-1.0)
Similarity ≥ this value = modified (0.0-1.0)
Hours before marking pending drafts as discarded