How to bulk generate images with ChatGPT

Set up your first ChatGPT batch, scale it to 100 prompts, and recover when a run stops partway.

Updated October 19, 2026 ChatGPT guide

Download for Google Chrome

This guide walks through AutoGem on chatgpt.com, from folder permissions to a finished run. For examples of what it can do, see AutoGem for ChatGPT.

Start with setup, jump to preparing a 100-prompt batch, or read how to review results and handle interruptions.

Set up a ChatGPT image batch

Install, sign in, add your prompts, and run a short test.

  1. Add AutoGem to Chrome

    Install from the Chrome Web Store, sign in to AutoGem, and open the side panel. Choose a download folder during setup or in Settings, then allow Chrome to write to it.

    AutoGem listing on the Chrome Web Store with the Add to Chrome button

  2. Connect to your ChatGPT tab

    Open chatgpt.com and sign in with an account that can generate images. Leave the tab open. AutoGem works through it, so you don’t need an API key.

    AutoGem side panel docked beside chatgpt.com with an image prompt queue prepared

  3. Build the bulk queue

    Paste one prompt per line, or separate longer prompts with a blank line. Set the aspect ratio and any reference images, then click Add to Queue. Check the rows before you start.

    AutoGem prompt box holding a list of ChatGPT image prompts above the queue

  4. Run a sample and check the downloads

    Start with a few rows. AutoGem sends each one in order and saves the images. Keep the side panel open so files reach your folder, and check both the queue and the folder.

    AutoGem queue rows listing each ChatGPT image prompt with its status

Get your prompts ready

For a storyboard, write each scene in full. Repeat the subject, style, and framing in every prompt instead of writing “same as before.” Add reference images where you need them, and check that the sample scenes match.

For a quick first run, download our three sample ChatGPT prompts, paste them in, and check the results before you add more.

Set the aspect ratio and attach references before you add the prompts. The model treats these as guidance, so check the size and subject of each image.

Check results and recover from interruptions

Check the images, then the files

Match the done and failed rows against your prompt list. Look for missing subjects, inconsistent details, or bad framing. AutoGem keeps each generated image even if saving to your folder fails. If a file is missing, find it in the extension, read any export warning, and restore folder access in Settings.

Continue a paused run with Resume

Pause lets the current image finish and holds your place. Resume picks up from there. Stop also lets the current image finish, but the next run after Stop, or after a finished run, starts from the first item again, including completed rows.

Retry only what failed

Read the error and look at the ChatGPT conversation before retrying. A rejected prompt fails on its own while the other rows keep going; a frozen ChatGPT page pauses the queue. Put the prompts you want to retry in a short new queue so you don’t rerun the whole list.

Collect the finished run

When a run finishes, Download ZIP collects all its images. Keep the prompt list with the files, and look over every image before you use it.

Related

See AutoGem for ChatGPT for product examples, or compare bulk image generation extensions.

Start your next ChatGPT batch on autopilot

Install AutoGem, pick a folder, and run a few ChatGPT prompts to see how it works.

Download for Google Chrome