Feature Drop: Image Upload Now Available

Ever wanted to show your AI assistant a screenshot, diagram, or photo? Now, you can upload images directly into your Omnifact conversations, enabling rich visual interactions with our AI assistants.

How to Get Started

Getting started is simple:

1. Enable "Image Input" in your organization's Experimental Features
2. Click the attachment button in any chat to upload an image
3. Select your image and watch the AI analyze it

We support PNG, JPEG, WebP, HEIC/HEIF, AVIF, BMP, GIF, ICO, SVG, and TIFF formats, and images up to 20MB. Images are automatically optimized for efficient processing while maintaining quality.

What You Can Do

Upload an image and ask your AI assistant to:

• Describe what it sees
• Answer questions about the content
• Analyze charts, graphs, or diagrams
• Help you understand complex visuals

Current Limitations and Future Improvements

We want to be transparent about our current capabilities and what's coming next. Here are some limitations we're working to address:

Generated Images vs Uploaded Images

Currently, there's a distinction between AI-generated images and uploaded images:

• Uploaded images work seamlessly - the AI can see and discuss them
• Generated images are stored separately and can't be directly referenced for editing

What This Means for Users

If you want to modify an AI-generated image, you'll need to download and re-upload it rather than just referencing the original. Also, generated images won't automatically appear in follow-up conversations (though the AI remembers the conversation about creating them).

We're actively working to unify this experience so all images work the same way.

Important Privacy Note

Your images are processed securely and never used for training. However: as this is an experimental feature, even if you have the Omnifact Privacy Filter™ enabled, images currently do not pass through the filter. Please enable this feature with caution when working with sensitive data.

Looking Forward

The limitations we mentioned are temporary - we're working hard to make image interactions even smoother. We're committed to continuously improving this feature based on user feedback and technological advancements.

We're excited to see how you'll leverage this new capability! Share your experiences with us to help shape future updates.