A Comic Generator based on GPT-40 and GPT Image 1

This set of Python scripts lets you generate a PDF Comic, complete with Images and description of each image, by starting off from one single initial "prompt".

The Code uses GPT to generate follow up scenes, and GPT Image 1 to generate an image for each scene described.

Finally, the scripts assembles the images together with their scene descriptions in a PDF Comic, featuring 4 Panels each page, a title, a footnote and a page-number in the footer.

1. You need an OpenAI API Key.
2. This program requires an .env file which is not provided in the repository.
   • You have to add it on your own within the main folder.
   • You have to add these contents and swap YOUR_OPENAI_API_KEY with the actual OpenAI API KEY:

   OPENAI_API_KEY=......
   MODEL=gpt-4o
   MAX_TOKENS=200
   TEMPERATURE=0.7
   FREQ_PENALTY=0.5
   PRES_PENALTY=0.7
   

   • You can, of course, alter the parameters where it makes sense. Please refer to the OpenAI DOCs for more details about what each of these settings do
3. You need the Python Libraries specified in the requirements.txt installed on your machine, and of course, Python (at least 3.11)

1. Download this repository and add your .env file to the main folder.
2. cd into the repository in a Terminal.
3. Run python3.11 main.py (if you use another Python handle, make sure to use your specific handle).
4. Complete the prompts:
   • When asked to Enter comic idea or theme:, enter the general idea for your comic (Example: An interstellar detective's adventures).
     • This will be used by GPT to generate the first image prompt, and from it, further prompts are consequtively generated.
   • When asked to Enter Artstyle, enter an artstyle.
     • Examples are provided in the input prompt.
   • When asked to Enter amount of Panels to generate:, enter a number that represents the amount of total panels in your comic.
     • Averagely, about 50 panels (50 images and 50 descriptions) cost 1 USD of API usage.
     • The program will put maximally 4 panels on each PDF page.
5. Let the program generate all panels. This can take a while.
6. Once finished, the terminal will show the message All panels have been generated. Run comic.py to finish your Comic!
7. Now run python3.11 comic.py
   • When asked to Enter your Comic Title, enter the Title you would like your finished comic to have.
8. Let the program finish and find your new Comic in the main folder, named comic.pdf. This should be pretty fast.
9. Cleanup for a new run:
   • Delete all contents in the logs.json file
   • Delete all images in the imgs folder
   • Delete the comic.pdf file

• You can alter several strings (prompts, credits) in the main.py file

• It is planned to add a GUI to the whole process, however this is not yet done. Contributions welcome!
• It is planned to make the images download asynchroneous, to speed it up a bit. ✅
• It is planned to add support for newer GPT models and once it is out