Amplify: Chat Interface

How to use the chat interface (center panel), select an AI model and its settings, and upload files to a conversation.

Using the Chat Interface: Start a new conversation

Amplify's chat interface allows you to choose various AI models, select certain AI settings, and attach files to a conversation. In this article, we will review:

1. The chat interface
2. How to choose a model and AI settings
   1. Custom instructions
   2. Temperature
   3. Response length
3. How to add files to a conversation
4. How to connect an AI Assistant to a conversation
5. Sparkle menu

Select a Model

1. Navigate to the "Model" dropdown in the center panel, and choose from the available models.  Amplify supports a variety of models for different use cases. AI models may change over time. On the right-hand side of the model selection are a number of icons that will connect you to settings or information.
   1. There are two rows of icons:
      1. From left to right, the top row contains cloud storage, temperature and response length settings (see below), and a legend.
         1. The default setting for cloud storage is disabled. This keeps your conversations and files attached to the personal device and the web browser you are using. If you wish to share across multiple devices, you will need to click on the cloud icon to remove the strike through, and activate cloud storage.
      2. From left-to right, the second row of icons give you information regarding the model you have chosen, including whether or not an image can be used in your prompt, general idea of model cost, and the average token output.
   2. For more tips on what model is best for what tasks you can also hover your cursor over each option with the dropdown menu open to read a short description.

Custom Instructions

Custom instructions in AI refer to predefined guidelines or parameters that shape how an AI system behaves, responds to user inputs, and processes information.  You can keep it Default, or use specialized assistants like the PowerPoint Assistant or Visualization Assistant to tailor responses to your needs.

Temperature & Response Length

When starting a new conversation in Amplify, you can adjust the Temperature and Response Length settings to control how the AI generates responses.

• Temperature (0.0 – 1.0): Controls creativity and randomness. Higher values (e.g., 0.8) increase randomness, while lower values (e.g., 0.2) make outputs more deterministic. The default is set to 1.0 for balanced behavior.
• Response Length (0.0 – 6.0): Adjusts how concise or verbose the model's responses are. Higher values allow for longer responses, with the default set to 3.0 for average length.

File Upload: Adding Existing & New Files

In Amplify, you can upload files to provide context or data for the AI to analyze and better integrate relevant materials into your conversations. There are two options:

• Add New Media: 
  • Click on the Add button in the chat or task window to upload files to Amplify.
  • Select the files from your device. Amplify supports various formats, including .pdf, .docx, .csv, and more.
• Add Existing Media:
  • You can access and reuse previously uploaded files.
  • You can also use the available filter options to sort by Name, Created Date, or Type for easily finding the document that you want to use.
  • The AI can reference the contents of uploaded files to enhance its responses.

Adding An Assistant

You can add a pre-made AI assistant to enhance your chat experience.

• For more information about how to create an assistant please refer to the Creating An Assistant section.

Message Input & Sending Your Message

Once you have your setting where you want them you can type your message into the chat box at the bottom of the screen. When you are done you can hit the send button to submit it to the AI.

Sparkle Menu

So what is that little sparkle found in your conversation panel? The sparkle symbol can be moved and therefore found anywhere on your Amplify chat interface. Clicking on it will reveal the following menu.

1. RAG mode: With this turned on, Amplify will try to pick out what it considers relevant information from any provided text your provide it in the prompt.
2. Artifact creation: Turning this on allows Amplify to create an Artifact space (a collaborative space typically used for coding or longer-form work). This allows you to edit Amplify's output in a separate panel.
3. Filter relevant information: When turned on in a conversation, Amplify will attempt to filter out what it considers the most important part of your conversation, and reuse those parts to help answer your prompts.

Questions or Issues?

For any questions or issues related to Amplify, please submit a ticket by selecting the Report Issue button found on the Amplify GenAI Issue page located in the ITS service catalog, under Teaching, Learning, & Research.

Acknowledgement

Initial documentation structure, text, and images were provided by the Amplify open source GitHub repository (Vanderbilt, 2025, and 40AU), and parts have been modified by Carleton College.