AI Prompt Compiler Bot

Diving deeper into AI, I wanted to find a way to streamline my video generation process by using a template. I built a bot that allowed me to fill out a form that would compile data that could be placed into an AI generator. You often get better results by including lots of detail so having a template to use allows me to focus on writing details vs. figuring out what to include.

Below is the process I used to create my app and put it online for others to use.

PYTHON CODE GENERATION

With my simple app idea, I used Chat GPT to describe what I wanted the app to do and had it generate the code for me. This was really easy for it to do, and after some tweaks, I was able to get to a place that felt right. I was able to make 95% of the changes through ChatGPT and had very little actual code edits on my end.

LOCAL TERMINAL

Using the generated code I could open the terminal and run the saved file locally to test it. Using this method I was able to make tweaks super fast, such as updating questions, reformating, etc.

FORM

The app is basiclly a set list of fields. You input your descriptions in each and let the program compile all the answers into a list. From there I included a line at the beginning that tells the AI generator to do two things:

1. Merge the following answers into a single cinematic, descriptive text block, optimized for AI platforms
2. Create a video based off the result.

HOSTING ONLINE

After adding new features and cleaning up the overall visual style (from Gemini's suggestions), I had Gemini convert the code to use Streamlit instead of tkinter logic.

Using Jules from Google, I was able to connect to my Github account, create a repository, and publish my code.

From there I created a streamlit account, connected it to GitHub, and created an app on Streamlit.

Now the app is hosted and can be used by anyone.

EXAMPLE GENERATED PROMPT AND VIDEOS VIA GEMINI

A muscular 40-year-old Italian knight, clad in heavily worn silver armor covered in scratches, dirt, and battle scars, grips a sword with both hands as he runs through a dense winter forest at dawn. His long blond hair spills out from the back of his helmet, and through the narrow slit of his visor we catch flashes of his piercing blue eyes, focused with determination. The camera moves handheld, as if running alongside the group of knights, bouncing slightly as it dodges trees, leaps over fallen logs, and ducks under low branches. Dawn light cuts through the dead trees, casting a cool blue hue over the snow-covered ground while warm shafts of light break through the forest canopy, glinting off the knight’s dented armor. The air is hazy with morning mist, and tiny particles of frost and snow catch the light, creating a shimmering effect. Motion blur enhances the frantic energy, while deep focus keeps the entire forest and the racing knights sharp. The scene is tense, chaotic, and cinematic — shot in 35mm style, 50mm lens, with heavy grain and natural low-light exposure, as if filmed by David Fincher and Jeff Cronenweth. Framing keeps the knight to the left side of the 16:9 frame as the forest blurs around him, immersing the viewer in a cold, gritty, hyper-real medieval chase sequence.

Try it out: