Skip to content
HTML/CSS to ImageDocs

Precisely capture specific elements from your webpage

The selector parameter allows you to capture specific elements from a webpage by providing a CSS selector. The API will automatically crop the image to match the dimensions of the selected element.

Consider this HTML structure:

<div class="margin">
<div class="content">
This is an example
</div>
</div>

To capture only the inner content div, set the selector parameter:

{
"html": "<div class=\"margin\">...</div>",
"selector": ".content"
}
Example of using a CSS selector to crop an image

To select elements by class name:

<!-- HTML -->
<div class="profile-card">...</div>
<!-- Selector -->
.profile-card

To select a unique element by ID:

<!-- HTML -->
<div id="header">...</div>
<!-- Selector -->
#header

To select nested elements:

<!-- HTML -->
<div class="container">
<section class="content">
<article class="post">...</article>
</section>
</div>
<!-- Selector -->
.container .content .post

To select elements with multiple classes:

<!-- HTML -->
<div class="card premium featured">...</div>
<!-- Selector -->
.card.premium.featured

For more complex selections, you can use:

Selector Example Description
Child parent > child Selects direct children only
Nth Child div:nth-child(2) Selects specific child elements
Attribute [data-type="premium"] Selects elements with specific attributes
Combinators header + .content Selects elements that follow others
  1. Be Specific: Use precise selectors to ensure you capture exactly what you need
  2. Test First: Verify your selector works in the browser before using it in the API
  3. Consider Dynamic Content: Allow time for JavaScript-rendered content using ms_delay if needed
  4. Unique Identifiers: When possible, use IDs for more reliable selection

If your selector doesn’t match any elements, try:

  • Verifying the selector syntax
  • Ensuring dynamic content has loaded (use ms_delay)
  • Checking for typos in class/ID names

If the wrong element is captured:

  • Make your selector more specific
  • Use unique identifiers when possible
  • Check for duplicates