Skip to main content
Home > “Admin Menu” > “Settings”
The system settings of your VARIOS AI instance

​
General

Max. costs per month in $
  • Set an upper limit for your maximum monthly costs
Proxy server address and port (e.g., “203.0.113.0:8080”)
  • Optional: Set up a proxy server, define exceptions (bypass list), and store the proxy certificate.
Default AI model for generating title and summary
  • Select the AI model you want to use for generating your chat titles and chat summaries
Default AI model for new assistants
  • Select the AI model that is preselected by default for a new assistant
Default AI model for generating embeddings
  • Select the AI model you want to use for generating your embeddings
Chats retention period (in days) (in days):
  • Infinite -> Chats are not deleted automatically
  • Delete after the retention period -> If users do not delete chats beforehand, they will be automatically deleted after the specified number of days
  • Do not delete before the retention period expires -> If users delete chats in advance, final deletion will only occur after the specified number of days
Retention period for log files (in days)
  • Set how many days your log files are retained
Knowledge base search mode
  • Define the search strategy for the embedding model. Best practice: Hybrid approach combining vector search and full-text search.
External service monitoring interval
  • Enable service monitoring and set the default check interval (in minutes) for our test tokens.
Percentage range for the yellow/warning indicator in the model monitoring dashboard
  • Define the percentage range for the yellow warning indicator in the model monitoring dashboard.

​
Privacy

​
PII Data Processing Mode

Choose how detected personally identifiable information (PII) should be handled before being passed on to the AI model. In anonymization, the original data is irreversibly removed; in pseudonymization, the PII values are substituted with placeholders before processing and automatically reinserted into the resulting model response.

​
Sensitivity Labels for Files

Sensitivity labels are a Microsoft feature that allows you to classify and protect files based on their sensitivity. With this feature, you can prohibit the use of classified files.

​
DLP Entities

Define which sensitive data the DLP module should identify and monitor:
  • User first name, last name
  • Credit card
  • IBAN
  • Cryptocurrency
  • Phone number
  • Email address
  • Person
  • IP address
  • NRP (Nationality, Religion, Politics)
  • Location

​
Deny List Entries

Add regex (regular expressions) with descriptions that should be explicitly filtered out by DLP.

​
Allow List

Add terms that should explicitly not be filtered out by DLP.

​
Permitted URL Protocols

Add your own protocols in addition to the standard protocols (http, https, mailto, tel). Data protection text
  • Define the Data protection text that is displayed when registering in the instance. This must be confirmed by every user before using VARIOS AI.

​
Theme Settings

Customize the appearance of your VARIOS AI instance to match your corporate branding.
Use PNG files with transparent backgrounds for the logos so they blend seamlessly into both color modes.

​
Advanced settings

Instance system prompt
  • This system prompt is applied to all chats in this instance. It defines the AI’s basic behavior in the instance.
Best-practice system prompt
  • This system prompt is used as a template for new prompts. It provides best practices for creating effective prompts.

​
System Prompt Generator

  • Define the model for the system prompt generator
It is recommended to use a high-quality AI model here.
Meta Prompt for Generation and Editing
  • This prompt is used for the generation and editing of system prompts with the system prompt generator.

​
Import Prompt Library

  • Define whether the prompt library should be imported
Reset Error Counter and Start Import
  • Restart the import of the prompt library

​
Expert Settings

The Expert Settings tab is only displayed if you have enabled the Expert View in your account settings.

​
Hybrid Search Weighting

Configure the weighting between vector search and full-text search for the hybrid search mode. The sum of both weights should equal 100%.
An even distribution (50/50) is a good starting point. Increase the vector weighting for semantically demanding searches or the full-text weighting for exact keyword matches.

​
Chunk Configuration

Configure how documents are split into sections (chunks) for processing by the embedding model.
Changes to the chunk configuration may require existing documents in the knowledge base to be re-embedded. Depending on the amount of data, this can take time and incur costs.

​
System information

Instance status Instance version Release channel
  • Extended Support Release (ES): Focus on stability and reliability
  • Monthly Release: Focus on innovation and feature growth
Account ID
  • The account ID of your instance
Enable support access
  • Enables temporary access for the VARIOS AI support team. The team can review the settings and configuration of your instance to assist with issues. User chats are not accessible.
Only enable support access when needed (e.g. when submitting a support ticket) and disable it again afterwards.
License key
  • The license key of your instance
System update System status:
  • Is your system up to date or is an update available
Change version
  • Update or downgrade your system to a different version
Create a backup of the server before updating the system.