Skip to main content

​
Extend AI Capabilities

Customize what your AI assistant can do. The Assistant settings page lets you toggle advanced features like Web Search and Image Generation, giving you control over the balance between capability and security. By default, all additional capabilities are disabled.
The Assistant settings page with toggles for Web Search, Image Generation, and Image Input.

​
Available Capabilities

Web Search and Browsing

When enabled, this allows Omnifact to search the internet and browse web pages. This capability is useful for providing up-to-the-minute information and gathering broader context for user queries that cannot be answered by the selected AI model alone.

Image Generation

Enabling this feature allows users to transform their ideas into visuals. You can configure the default image generation model for your team.

Image Input (Experimental)

This experimental feature allows users to add images directly into a chat to support multimodal interactions, such as asking questions about a chart or diagram.
This feature has significant privacy implications.Images uploaded via this feature do not pass through the Omnifact Privacy Filter™. While images are never used for training by the underlying model provider, any sensitive data contained within the images will be sent directly to the model.Please enable this feature with extreme caution and ensure your users are aware of the risk when handling sensitive data.

Audio Input (Experimental)

This experimental feature allows users to add audio files directly into a chat to support multimodal interactions with audio-capable models.
This feature has significant privacy implications.Audio files uploaded via this feature do not pass through the Omnifact Privacy Filter™. While audio files are never used for training by the underlying model provider, any sensitive data contained within the audio will be sent directly to the model.Please enable this feature with extreme caution and ensure your users are aware of the risk when handling sensitive data.

Inline Sources (Experimental)

Display source citations inline within the assistant’s response for better context and attribution.

​
Configure Default Image Model

You can choose the specific image generation model that your team uses.
  1. Navigate to Team Settings > Assistant.
  2. Locate the Image Generation section.
  3. Select your preferred provider and model from the dropdown menu.

​
Available Models

  • GPT-Image-1: OpenAI’s flagship model, known for its versatility and photorealistic results.
  • Nano Banana: A high-speed, creative model perfect for rapid iteration and stylized outputs.
  • Nano Banana Pro: The premium version for professional workflows, featuring higher resolution, intricate detail retention, and advanced composition control.