The default template for the chatbot AI agent processes purely text inputs, and cannot interpret attached files or images. However, if you wish to equip your chatbot with the ability to see file attachments, you will need to provide it with the appropriate tools.
Please note that expanding your AI agent's capabilities will result in higher credit consumption.
Currently, AI agents can only handle attachments during chats. Work with ticket messages that contain attachments is currently not supported. If you require this functionality, please do not hesitate to contact us.
Equipping your AI agent with vision and attachment tools
To enable your AI chatbot to see attachments in a chat, login to your FlowHunt account, navigate to My Agents, and edit your chatbot AI agent (default name "Chatbot for LiveAgent Integration"). In the flow builder, click on the AI Agent component, and in the Tools section, click "+ Add Tool".
Vision Tool
The Vision Tool is the first tool that enables your AI agent to see images attached in a chat, such as JPG or PNG files.
After selecting the tool, you can configure it. Make sure that the Prompt and Image URL parameters are set to AI decision mode. Optionally, you can specify the LLM model you want to use exclusively for this tool (e.g. gpt-4o, gpt-5.2, claude-4.5-haiku, gemini-3-pro-preview, etc.). Click "Add with Config" to finish and add it to your AI agent's tools.
File Attachment Retriever
The File Attachment Retriever tool is the second tool that allows your AI agent to interpret attached files in a chat. This tool can read the content of files such as PDF, DOCX, TXT, and others, excluding image files.
After selecting the tool, you can adjust its settings.
- Max Tokens - Define the maximum number of tokens the AI can use to generate an output with this tool; or allow the AI to decide how many tokens it wants to spend.
- Strategy - Select whether multiple attachments should be thoroughly reviewed sequentially one by one, or if all should be treated with equal importance; or alternatively, let the AI decide the strategy.
- Apply OCR - Choose whether you want to apply OCR (optical character recognition) to enable your AI agent to see also images inside attached files or not; or let AI to decide if it wants to use OCR.
Click "Add with Config" to finish and add it to your AI agent's tools.
That's it. With these two tools, your AI chatbot will be able to see images and files attached by users during chats and respond appropriately based on the information in the attachments.
To restrict the types of files that users can upload during a chat, navigate in LiveAgent panel to Configuration > Security > Settings, click the "Change" button in the "General" section, and enter only allowed file types into the "Allowed file types" field.