Chat Features

Overview

The chat interface is the primary way to interact with your agents. Send messages, attach files, and watch your agent use its connected tools to complete tasks in real time.

Each conversation is saved as a thread. You can start new conversations, switch between existing threads, or clear history from the menu in the header.

File Attachments

Click the paperclip icon to attach files to your message. Your agent can read and analyze file contents — useful for processing spreadsheets, reviewing documents, or extracting data from PDFs.

Image Generation

Ask your agent to generate images and it will create them using AI. Images are displayed inline in the chat and can be downloaded. Useful for creating social media graphics, diagrams, or illustrations.

Code Execution

Agents can write and execute Python code for data analysis, calculations, file processing, and more. Code runs in a sandboxed environment and output is displayed below the code block.

Web Browsing

All agents can browse the web to research information, visit URLs, and extract content from pages. Ask your agent to "look up" or "research" something and it will use the browser tool.

Presentations

Agents can create PowerPoint presentations from scratch. Describe the content and structure you want, and the agent will generate a downloadable .pptx file with formatted slides.

Tool Visibility

When your agent uses tools, you'll see each step in a collapsible chain-of-thought view:

• •The tool name and app icon are shown for each action.
• •A spinner indicates when a tool is running.
• •A checkmark or X shows success or failure.
• •Elapsed time is displayed for each tool call.
• •A brief result summary is shown under each step.
• •Parallel tool calls are grouped and displayed together.