Skip to main content

Images

You can attach an image to the chat. This is especially useful for frontend development and UI testing.

To do this:

  1. Click the plus icon in the chat
  2. Select the Image(s)... option
  3. Choose the image you need
  4. The image appears as an attachment to the chat
  5. When you send the message, the agent receives the image and can see its content

Not all models support images. If your main model works with text only, Explyt can pass the image to a subagent whose model can read it.

Images in subagents​

A text-only model can pass an image to a subagent and get back a description or the extracted text. This is useful when you need to analyze a mockup, read an error message on a screenshot, or check an interface without changing the main model.

Create a subagent profile, choose a model with image support, and describe what the profile is for. The main agent uses that description to decide when to hand it a task.

If the main model cannot read the image itself, it receives the file path, passes it to a subagent, and uses the result the subagent prepares.

Pasting via Ctrl+V and drag-and-drop​

Images and files can be added to the chat via Ctrl+V (paste from clipboard) and drag-and-drop. Just take a screenshot and press Ctrl+V directly in the chat input field.

If the selected model does not support images, the main model can pass the file to a suitable subagent.

This is useful when you need to quickly show the agent an error on screen, a UI mockup, or the result of a program run without the extra steps of saving and attaching a file.

Automatic image compression​

Images attached to the chat are automatically downscaled before being sent to the model. By default, Explyt limits the larger side of an image to 1024 px. If neither the width nor the height exceeds the threshold, the image is left unchanged. The minimum allowed value is 32 px.

Leave the value at 1024 px for ordinary interface screenshots. Raise the threshold if the model needs to make out small text, diagram elements, or mockup details. Lower it for large overview screenshots without fine detail, to reduce token consumption. Turn automatic compression off if it is important to pass the image at its original resolution; this increases the amount of context used.

To change the setting:

  1. Open Settings | Tools | Explyt Agent.
  2. Enable or disable Enable automatic image compression if max dimension >.
  3. If compression is enabled, specify the maximum size of the image's larger side in pixels, no less than 32.