Advanced Settings
The Advanced Settings section contains three main components: Chunking, Metadata, and Video Settings.
Figure: Advanced Settings interface (default)
10.1. Chunking
Figure: Chunking interface
Field/Setting | Description |
|---|---|
Reranker (Toggle Switch) | Enables or disables the Reranker service |
Reranker Model | Specifies the Reranker model to use |
Show Reference Link in Answer (Toggle Switch) | Allows the AI agent to include source links in its responses |
Enable Suggested Questions (Toggle Switch) | Displays follow-up questions to guide the conversation |
Number of Suggested Questions | Defines how many suggested questions the agent generates |
Chat History Aware (Toggle Switch) | Enables or disables the agent's ability to remember previous messages to maintain context |
Number of Messages to be Aware Of | Defines how many prior messages the agent considers when generating responses |
Max Output Tokens | Sets the maximum number of tokens the AI agent can generate in a single response |
Vector Store | Specifies the database to use |
Embedding Model | Identifies the specific embedding model to use |
Chunk Size | Determines the maximum size (in tokens or characters) of document pieces before processing |
Chunk Overlap | Sets the amount of shared content between consecutive chunks |
Top K Value | Limits the search to a fixed number (K) of the most relevant results when generating responses |
Top P Value | Controls response diversity using nucleus sampling |
__TABLE_CAPTION__Table: Fields and Settings under Chunking__TABLE_CAPTION__
10.2. Metadata
The Metadata section provides essential descriptive context to organize and manage the agent's data sources effectively.
Field/Button | Description |
|---|---|
| Creates new metadata |
Name | Defines the metadata name |
Cancel | Cancels metadata creation |
Create | Creates the specified metadata |
Save Changes | Saves metadata updates |
Table: Fields and Buttons under Metadata
Figure: Metadata interface
Figure: Create Metadata dialog box
Figure: Metadata interface with filled details
Figure: Success message after clicking "Save Changes"
10.3. Video Settings
Figure: Video Settings interface
Field | Description |
|---|---|
Video Description Model | Selects the AI model used to analyze video content and generate descriptions |
Seconds per Frame | Defines the time interval (in seconds) between video frames that the AI analyzes |
Description Prompt | Instructs the agent on how to analyze video frames and what details to include in descriptions |
Transcription Prompt | Guides the AI on how to transcribe audio from videos into text format |
__TABLE_CAPTION__Table: Fields under Video Settings__TABLE_CAPTION__