Model Management
The model management provides a clear interface for configuring AI models for chat and embedding applications.Create New Models
1
Open Model List
Navigate to the model management in the admin interface and choose between “Chat Models” and “Embedding Models”.
2
Add New Model
Click the “New Model” button in the upper-right corner.
3
Enter Basic Data
Fill in the fields:
- Image: Choose a meaningful image for the model
- Title: A descriptive name for the model
4
API Configuration
Configure the API settings:
- API: Select the appropriate API type
- Credentials: Select the associated credentials
- Model Name: The specific model name
5
Configure Costs and Limits
Set token costs and context limits:
- Cost Input Tokens: Cost per input token
- Cost Output Tokens: Cost per output token
- Maximum Tokens per Context: Maximum number of tokens in the context
6
Behavior Settings
Configure model behavior:
- Temperature: Model creativity (0-2)
- Top P: Nucleus sampling parameter (0-1)
- Reasoning Effort: Cognitive effort of the model
- Streaming: Support for streaming responses
7
Security Settings
Configure DLP (Data Loss Prevention) settings:
- User Identification: Adds a unique identifier to the API request
- Allowed MIME Types: Protection against certain file types (copyable)
- GDPR Protection: Protection of personal data
- Sensitive File Protection: Protection of confidential documents
- Enterprise DLP: ICAP server integration
8
Save
Click “Save” to create the model.
Available API Types
OpenAI Chat API (/chat/completions)
Standard OpenAI Chat Completions API for text generation.OpenAI Assistant API (/assistants)
OpenAI Assistant API for structured assistant functionality.OpenAI Responses API (/responses)
OpenAI Responses API for special response formats.IONOS API (/chat/completions)
IONOS-compatible Chat Completions API.OpenAI Chat API without HTTP Proxy
OpenAI API without proxy support for direct connections.DLP Security Settings
Basic DLP Settings
GDPR Protection
Detects personal data in the chat and encrypts it before transfer to the AI. Protection level for GDPR-compliant data processing. Available Options- enabled: GDPR protection is always enabled and cannot be disabled.
- optional: GDPR protection is enabled by default but can be disabled by the user
- disabled: GDPR protection is disabled and not applied.
Sensitive File Protection
Blocks the upload of documents that have a confidential classification. Protection level for confidential documents. Available Options- enabled: Upload of confidential documents is blocked.
- disabled: Upload of confidential documents is allowed.
Enterprise DLP Settings
DLP Interface (ICAP)
Sends texts and files to the configured ICAP server for inspection. Activation of enterprise DLP functionality. Available Options- enabled: ICAP server is used for all content.
- disabled: ICAP server is not used.
Manage Models
List of Models
The overview shows all configured models with:- Icon: Visual identifier of the model
- Name: The specified name of the model
- Type: Chat or Embedding model
- Actions:
- Edit (pencil button)
- Delete (trash button)
Model Types
Chat Models
Models for text generation and conversations.Embedding Models
Models for text vectorization and semantic search.Best Practices
1
Model Selection
Choose the appropriate model based on your requirements and budget.
2
Cost Optimization
Configure realistic token costs for accurate cost tracking.
3
Security
Enable DLP protection for sensitive data and compliance requirements.
4
Performance
Set appropriate token limits for optimal performance.
5
Testing
Test new model configurations in a safe environment.
Info: Models can only be used with credentials of the corresponding provider type.
Note: Fields marked with an asterisk (*) are required and must be filled out.