User guides

Use attachments in Pixso AI

Add a document or image to a Pixso AI conversation to provide requirements, page copy, brand guidance, or a visual reference. Then select the design you want to edit and explain how AI should use the attachment.

For example, a requirements document can supply the content for a page, a brand guide can explain its colors and typography, and a reference image can show the layout or style you want.

Upload an attachment

  1. Open your Pixso design file.
  2. Select AI at the top of the left sidebar.
  3. Select the design or layer you want to edit.
  4. Use the attachment icon in the conversation input.
  5. Choose a file from your device.
  6. Describe the requested edit and send the message.

An uploaded attachment appears in the prompt area. To remove it before sending the message, use the attachment's remove control.

Use documents

Documents can provide requirements, approved copy, design instructions, or brand conventions. Choose a document format accepted by the attachment file picker.

You can ask AI to:

  • Add page sections described in a requirements document.
  • Replace placeholder text with approved copy.
  • Adjust colors, fonts, and corner radii using brand guidance.
  • Apply review comments to the selected design.

For a long document, name the section or requirements that matter to the current edit. This keeps the task focused on the information needed for that part of the design.

Use images

You can attach an interface screenshot, visual reference, brand asset, or product image. Choose an image format accepted by the attachment file picker.

Explain whether the image is a reference or an asset to place in the design:

  • Visual reference: use its layout, colors, typography, or visual style to guide the edit.
  • Design asset: use the image itself in the design, such as a banner background or a content illustration.

AI may interpret small text, icons, or complex details differently from the original. A very large image may also be compressed. Use a clear reference focused on the relevant area, with readable text where text matters.

Give each file a role

When you add several files, explain what each one is for. For example:

Use product-requirements.pdf to determine the page content, use homepage.png as a visual reference, and replace placeholder text with the copy in product-copy.txt.

This is more specific than “update the design using these files,” which leaves AI to decide how their contents relate to one another.

Distinguish a reference from an asset

For a visual reference, describe the features to follow:

Use the attached image's color palette, typography, and card layout to refine the selected page. Do not copy its text or icons.

For an image that belongs in the design, state where to place it:

Use the attached image as the top banner background. Preserve its proportions and keep the heading readable.

A general request such as “make it like this image” does not tell AI whether to copy the layout, borrow the style, or place the image itself.

Keep the input focused

Attach only the files needed for the current task. A relevant requirements document and one reference image are often easier to work with than a large set of unrelated materials.

Split a long document by section when appropriate. Crop a complex screenshot to the area you want to discuss, while keeping enough context to understand the layout. If a file is too large to upload, reduce its size or split it into smaller files.

Work through complex designs in sections

For a complex page, complete one section before moving to the next. Start with a bounded request:

Use the attached requirements to update the top navigation, including the logo, menu, and user area. Leave the other sections unchanged.

After reviewing that result, continue:

Now update the data overview. Add the required statistics cards and make their spacing, corner radii, and text hierarchy consistent.

This gives each request a clear scope. For a multi-page design, work through the pages individually instead of asking AI to redesign every page and module in one message.

Check the result and resolve upload problems

Review the generated text, layout, images, and design details against your source material. AI can misread an attachment or apply a requirement to the wrong part of the design.

Use files you have permission to share. Remove passwords, identity documents, confidential business information, and unrelated or outdated content before uploading.

If an attachment cannot be read, try a supported format, a smaller file, or a shorter section of the document. If an upload fails or the conversation stops unexpectedly, keep the session ID and details of the problem when contacting Pixso support.