How to configure the conversation tab
The conversation tab controls which media types the agent processes and how it formats responses.
Available options:
- Transcribe audio: the agent transcribes received audio before processing a response. Adds processing cost.
- Interpret image: the agent analyzes and describes images sent by the user. Adds processing cost.
- Read files: the agent reads file content such as PDFs. Set a character limit to avoid very long documents. Adds processing cost.
- Split response message: instead of sending the full response at once, the agent breaks it into parts.
- Character limit per message: sets the maximum size of each part when message splitting is active.
- Audio response: the agent replies with audio. Choose the voice type and define behavior instructions for this mode.
- Default message for invalid content: text sent when the user sends something the agent cannot recognize. Example: "I cannot understand this type of message."