5 minutes read

Level: Advanced

Last Updated: December 2025

GPT Integration for Image Recognition

This feature helps to recognize and understand the contents of images shared by the end users. It is useful for a variety of use cases such as:

• Grading worksheets or assignments
• Reading and digitizing handwritten notes
• Checking skin conditions, wounds, or other health-related visuals
• Reviewing the condition of water, crops, or surroundings

With GPT’s image processing abilities, Glific makes it easy to automate tasks that involve looking at and understanding images.

To get the best results:

• Use clear and well-lit images
• Avoid blurry or low-quality photos
• Make sure the main content is visible and not blocked

---

How to Use in a Flow

Step1: Collect Image Input from the end user

• Create a send a message node directing users to send images as their response, based on their preference.
• In the Wait for response node, select Has image as the message response type. Also, give a Result Name. In the screenshot below, image is used as the result name.

Step 2: Add a Webhook Node to Process the Image

• By default, Function would be selected. Leave this as it is.

• In the FUNCTION field, select the pre-defined function parse_via_gpt_vision, from the dropdown.

• Give the webhook result name - you can use any name. In the screenshot example, it’s named gptvision.

---

Step 3: Add Parameters in Function Body

• Click on FUNCTION Body on the top right corner. You would see the following.

• Pass the parameters as shown in the below screenshot.

• Url: in this field pass the flow variable accepting the image response from the user (In the given example image is the result name).
• prompt: in this field pass the prompt, or instructions you wish to convey to the AI model towards processing the image input.
• Model: in this field pass the gpt model.

Step 4: Display the text response

• Create a Send Message node.
• Use @results.webhook_result_name.response to show the text response (In the given example gptvision is the webhook result name).

---

Sample Flow

Try this Sample Flow to test the GPT Vision integration.