How to Use GPTImage123 — Full Tutorial

Overview

GPTImage123 is an online creative workspace built around GPT-Image-2, OpenAI's latest image model. Instead of juggling API keys and command lines, you get a browser interface for turning ideas into finished visuals — plus a PPT generator for turning a topic into slides. This tutorial walks through every feature so you can get results on your first session.

If you just want to start, open the Image Generator, type a prompt, and press Generate Images. Read on for the details that make the difference between an "okay" result and a great one.

1. Creating an account and getting credits

1. Click Sign in in the top-right corner. You can register with email or a supported social login.
2. New accounts receive a small number of free credits so you can try generation before paying.
3. Every generation spends credits. Your remaining balance is always shown next to the generate button and on the Credits page.
4. When you run low, visit Pricing to add credits. You can choose a monthly subscription (credits refresh every month) or a one-time top-up pack (credits stay valid for a fixed window). Pick top-ups if your usage is occasional, and a subscription if you generate regularly.

Tip: credits are consumed whether or not you love the result, so it pays to write a clear prompt the first time. See the FAQ for prompt-writing techniques.

2. Text to Image

This is the core feature. Open the Image Generator and make sure the Text to Image tab is selected.

1. Choose a model. GPT-Image-2 is the default and recommended choice. You can switch to older models to compare quality or cost.
2. Write your prompt. Describe the subject, style, lighting, camera angle, and composition. The more specific you are, the more predictable the output.
3. Press Generate Images. Progress is shown while the model works. Results appear in the gallery below the form.
4. Download or iterate. Save the image, or tweak your prompt and generate again to refine.

A good starting structure for a prompt is: subject + setting + style + lighting + composition/aspect. For example:

A ceramic coffee mug on a wooden café table, morning light from a window on the left, shallow depth of field, warm color grade, photographic, 3:2 aspect.

3. Image to Image (using reference images)

Switch to the Image to Image tab when you want the model to work from an existing picture — for editing, restyling, or keeping a subject consistent.

1. Upload one or more reference images. These guide the model's output.
2. Write a prompt describing the change you want — for example, "put this product on a marble surface with soft studio lighting," or "redraw this character in a flat illustration style."
3. Generate. The model blends your reference with your instructions.

Use this mode for product photography variations, character consistency across a set, style transfer, and localized edits. If uploads fail, remove the affected image and upload it again.

4. PPT Generator

The PPT Generator turns a topic or outline into a slide deck.

1. Enter your topic, subject, or a rough outline.
2. Generate the deck and review the structure and content.
3. Export or refine as needed.

This is useful for first drafts of pitches, lesson plans, and internal presentations — start from the generated structure and edit rather than building from a blank page.

5. History and re-using past work

Every generation is saved to your History page. Use it to:

• Find and re-download images you created earlier.
• Review the prompts that produced your best results, and reuse them.
• Track how you've spent credits over time.

Treating history as a personal prompt library is one of the fastest ways to improve — when something works, save the prompt and adapt it.

6. Tips for better results

• Be specific, not verbose. One precise sentence beats a paragraph of vague adjectives.
• Name a style explicitly ("watercolor," "isometric 3D render," "35mm film photo") rather than hoping the model guesses.
• Iterate in small steps. Change one variable at a time so you can tell what actually helped.
• Use reference images whenever you need consistency in a character, product, or brand look.
• For text inside images, put the exact words in quotes in your prompt. GPT-Image-2 renders text well, but being explicit helps.

For a deeper dive into prompt engineering and model capabilities, read the FAQ. To understand how GPT-Image-2 compares to Gemini, Seedream, Midjourney, FLUX and others, see our image model comparison.

Troubleshooting

• "Not enough credits." Top up on the Pricing page.
• Upload failed. Remove the image and re-upload; check the file is a common format and not overly large.
• Result isn't what you wanted. Refine the prompt with more specific style and composition terms, or switch to Image to Image with a reference.
• Generation is slow. Peak times can add to the queue; a higher-tier plan offers priority.

Ready to create? Head to the Image Generator and start with a single, clear sentence.