Display images in agent responses
This guide explains how to enable image citations so your agent can search through uploaded images, reference them as sources, and display them directly in conversations.
Learn how to configure your agent for consistent image display.
Step 1: Enable AI Vision for uploaded images
AI Vision allows your agent to analyse and understand visual content. When enabled, the platform generates a description of each image that enables semantic matching between user queries and relevant images. To enable AI Vision:
- Click the three dots next to your agent’s name, then select Build.
- Click Add Source.
- Select File Upload.
- Select one or more image files (e.g., JPG, PNG, or diagram screenshots) and click on Advanced Options.
- Scroll down and toggle AI Vision in the upload window to enable intelligent image analysis.
- Click Add Files to confirm the upload.
Learn more: Enable AI Vision for uploaded images
Step 2: Activate image citations
Image citations allow your agent to display images directly in responses when they are referenced as sources.
- Go to your agent and click Personalize.
- Navigate to the Citations tab.
- Scroll down to the What to do with citation if it’s an image section.
- Click the dropdown menu and select Insert image instead of a numbered citation.
- Scroll down to Limit the height of image citation
- Select an option from the dropdown:
- Yes to adjust image height proportionately to chat area
- No to display images at full size
- Click Save Settings to enable the feature.
Learn more: Activate image citations
Persona instruction best practices
Avoid these instructions:
- "Show me the image" without relying on citations
- "Only show the image if..." (conditional logic)
- "If you find an image, embed it."
These instructions cause inconsistent output and broken links.
Use this recommended instruction:
"If you use an image as a source, always include an inline citation for that image."
Understanding image quotas
Your plan includes a monthly image quota—for example, 200 images per month. This refers to the number of images you can process and add to your knowledge base (ingestion), not how often images appear in responses.
For example, if you upload 50 product screenshots, you've used 50 of your quota. Those 50 images can then appear in hundreds of chat responses without consuming additional quota.
Related articles
- How AI Vision works
- Vision processing limits
- Enable AI vision for uploaded images
- Activate image citations
- Embed your photos in agent responses
Updated 24 days ago
