Hello, I think it’s absolutely great that ChatGPT can now create images with text and without any spelling mistakes! My suggestion: it would be fantastic if the bot could, for example, create an image with two people and display the text in speech bubbles (that way, I could use an image to tell a joke or make a comic strip)!
New “4o Image Generation” can do that, yes!
Prompt
Create a four-panel comic strip with clean lines and bright colors, in a modern cartoon style. Each panel features two people—Person A (a curious tech user) and Person B (an AI assistant avatar with a friendly look). Both characters have distinct looks and expressive faces. Include clearly readable speech bubbles with the exact text below:
Panel 1: Person A, surprised, asks: “Can ChatGPT write text in image?”
Panel 2: Person B points upward confidently and says: “Of course, look at the speech balloon.”
Panel 3: Person A, amazed, says: “Oh really, it’s my words!”
Panel 4: Person B smiles and waves: “Hey @marcuspc1981, Welcome to the community!”
Background: Simple, indoor tech setting with light shadows. Text must be perfectly legible in white speech bubbles, clean sans-serif font.
1 Like
Thank you, somethimes I have some mistakes with 3 ore 4 speaik boubles, but it works (for example with Jokes for a newspaper)
1 Like
That’s what I mean in paddle 6: it’s spelled “Küche” with an umlaut – Ü! ChatGPT still struggles a bit with German umlauts!
1 Like
Yes, it still struggles with non-English letters like “ö,” “ü,” “ç,” “ş,” “ğ,” and other characters used in various languages. OpenAI is probably working on improving this. The new ImageGen model is better than DALL·E when it comes to rendering text, but it’s still not perfect for all languages.
Prompt-1
Create a 6-panel cartoon-style comic illustrating a recipe for baked bell peppers (“Gebackene Paprika”). Use the Comic Neue font for all text to support German characters like umlauts (ä, ö, ü) and the Eszett (ß), and to match the comic style. Each panel should have a title in German and include the described characters, actions, and speech or text. Titles should be in rectangles at the top of each panel.
Panel Descriptions:
-
Top Left Panel
Title: Heute kochen wir: Gebackene Paprika!
A smiling cartoon chef in a white uniform with a green neckerchief holds a red bell pepper proudly. Background suggests a bright kitchen. -
Top Right Panel
Title: Vorbereitung
The chef slices red bell peppers on a cutting board.
Text box in the corner: Paprika vorbereiten
Speech bubble: “Jetzt schneide ich die Paprika in Streifen!” -
Middle Left Panel
Title: Marinade
The chef is mixing ingredients like oil, herbs, and garlic in a bowl.
Text box in the corner: Marinade zubereiten
Speech bubble: “Jetzt rühre ich alles gut um!” -
Middle Right Panel
Title: Backen
Sliced bell peppers are arranged on a baking tray being placed into an oven.
Text in the corner: Jetzt geht’s für 23 Minuten in den Ofen bei 200 Grad. -
Bottom Left Panel
Title: Anrichten
The chef is plating the baked peppers onto a serving dish.
Speech bubble: “Fertig! Dazu passt glutenfreies Brot mit Frischkäse oder Gurkensalat.” -
Bottom Right Panel
Title: Guten Appetit
The chef holds the finished dish proudly, with a smiling face character next to him giving a thumbs up.
Speech bubble: “Guten Appetit mit der Förde-Küche!”
Prompt-2
Next, proceed with the second image: a single-panel, cartoon-style comic with a tall aspect ratio. In this scene, a customer looks to tje chef and surprised and speaks in a mix of German and Turkish. You should write letters “ş”, “ö”, and “ı” correctly. The speech bubble should include the exact, unaltered text: “Aber ich möchte şiş köfte, cacık und ayran!”
1 Like