How to Use AI Agent in PhotoDirector

Q: Can I use more than one photo in a request?

Yes. Click Images in the request box to add up to 4 photos. Then type @ in your request to say which one you mean, as in Replace the handbag in @image1 with the @image2 .

Q: Where are my AI Agent images saved?

Images generated with credits are saved in My Creations . To save a version as a file, click Download and choose PNG, JPG, or TIF. Clicking New Chat clears the other images in the chat, so download what you want to keep first.

Describe the change or the picture you want, then keep asking until it looks right.

Updated Oct. 26, 2026 · 4 min read

In a hurry? Use the AI Agent quick guide instead.

PhotoDirector's AI Agent edits and creates images from a written description. Attach your own photo and say what to change, or describe a new picture, then keep asking for changes in the same chat. Use it for edits that would take several tools by hand, such as replacing what's behind a person or turning a spring scene into winter.

BEFORE

A woman in a straw hat and glasses smiles on a cobbled street with an old tram behind her — The original photo.

AFTER

The same woman and pose, with a sidewalk cafe, outdoor tables, and flower pots where the tram was — After asking AI Agent to replace the tram with a sidewalk cafe.

⚡ Quick Answer: How to Use AI Agent in 5 Steps

Open PhotoDirector. The AI Agent request box is at the top of the launcher.
Click Images in the box and choose your photo.
Describe the change you want and click the send arrow.
Click Yes to confirm the credits.
Click Download to save the result, or Use to keep editing it.

Steps to Edit Your Own Photo with AI Agent

Step 1: Open AI Agent

Open PhotoDirector. The AI Agent request box sits at the top of the launcher's Home screen, under How would you like to create?

The PhotoDirector launcher Home screen with the AI Agent request box at the top highlighted — The request box, on the launcher.

Step 2: Attach Your Photo

Click Images at the left of the box and choose the photo you want to edit. It appears in the box, labeled @image1. The box takes up to 4 photos.

The launcher request box with the Images button and the attached street photo labeled @image1 highlighted — Your photo, attached as @image1.

Step 3: Describe the Change

In the same box, say what to change and what it should become. This example uses Replace the tram behind her with a sidewalk cafe with outdoor tables and flower pots. Then click the orange send arrow at the bottom right of the box.

The launcher request box with the street photo attached, the cafe request typed in, and the send arrow highlighted — Your photo, plus the change.

Step 4: Confirm the Credits

The AI Agent window opens with your request in the chat. A message tells you how many credits the edit uses. Click Yes to go ahead. To skip this question next time, check Always allow credit usage first.

The AI Agent window with the cafe request in the chat and the Yes button of the credit message highlighted — Click Yes to make the edit.

Step 5: Compare Before and After

AI Agent replaced the tram and kept her pose, her hat, and her outfit. To check the change, click the compare button, the second icon at the bottom left of the canvas, and see Before and After side by side.

The street photo in Before and After views, the tram on the left and a sidewalk cafe on the right, with the compare button highlighted — The tram, then the sidewalk cafe.

Step 6: Download or Keep Editing

Click Download to save the photo as a PNG, JPG, or TIF file. Or click Use to open it in Guided & GenAI and keep editing it with PhotoDirector's other tools. To change it again with AI Agent, type another request in the box at the bottom left.

The edited street photo with a sidewalk cafe behind the woman, and the Use and Download buttons highlighted — Download it, or keep editing.

For ideas on what to ask, click Inspiration at the top of the chat. Each card shows a before-and-after example and its prompt, and Apply Prompt fills that request in for you.

The Inspiration window with the Replace, Remove, Enhance, and Restoration tabs and the Change Weather Apply Prompt button highlighted — Ready-made requests, grouped by tab.

The AI Agent page in the launcher has more ideas under Explore Idea. Hover over a picture to compare its before and after.

The launcher's AI Agent page with the AI Agent menu item and a hovered Explore Idea example highlighted — More examples on the launcher's AI Agent page.

Steps to Create an Image from a Description

Step 1: Describe the Image You Want

Without attaching a photo, type a description of the picture in the launcher's request box, with the subject, the setting, and the light. This example uses A golden retriever puppy sitting in a field of sunflowers, warm late-afternoon light, shallow depth of field. Then click the send arrow, and click Yes to confirm the credits.

The launcher request box with the puppy description typed in and the send arrow highlighted — A description, then the send arrow.

Step 2: Review the Result

The image appears on the canvas and in the chat. To take it further, open Final Touch under the canvas and choose Upscale for a larger version or Turn Into Video to animate it.

The generated puppy image on the canvas with the Final Touch menu open, showing Upscale and Turn Into Video — Final Touch: Upscale or Turn Into Video.

Step 3: Ask for Changes

The result is already in the request box, labeled @target, so your next request changes it. Type what to change, such as Add a red bandana around the puppy's neck, and click the send arrow. Then confirm the credits again.

The request box with the result labeled @target, the follow-up request typed in, and the send arrow highlighted — The result as @target, plus a follow-up request.

Each new version joins the strip under the canvas, after Original. Click a thumbnail to go back to it, and click Download when a version looks right.

The puppy with a red bandana on the canvas, with the version strip, Use, and Download highlighted — Every version stays in the strip.

The Result

The follow-up request added the bandana and left the puppy, the sunflowers, and the light as they were.

FIRST REQUEST

A golden retriever puppy sitting in a sunflower field in warm late-afternoon light — Created from the description.

FOLLOW-UP

The same puppy and field, now with a red bandana around its neck — After asking for a red bandana.

We tested these steps in PhotoDirector 365 version 18.0.10 on Windows.

Don't have PhotoDirector yet? Attach a photo, say what to change, and AI Agent edits it for you.

Download PhotoDirector Free →

FAQ About AI Agent

1. Can I use more than one photo in a request?

Yes. Click Images in the request box to add up to 4 photos. Then type @ in your request to say which one you mean, as in Replace the handbag in @image1 with the @image2.

2. Where are my AI Agent images saved?

Images generated with credits are saved in My Creations. To save a version as a file, click Download and choose PNG, JPG, or TIF. Clicking New Chat clears the other images in the chat, so download what you want to keep first.

3. How is AI Agent different from Quick Actions?

Quick Actions suggests standard edits for a photo, such as exposure or a color filter, and you fine-tune them with sliders. AI Agent follows your own description, so it can make changes no slider covers, like replacing an object or changing the weather.