AI Presentation Generator

Mar 03, 2025

---

Are you spending hours to create a presentation? Instead, you can let AI do the work for you.

Here is how you can code a presentation generator yourself. I included the coding prompt I used in the second part of this article.

• Use Cursor AI code editor to create a presentation generator based on the prompt I attached below.
• Sign up for AI APIs:
  • Set up a key to use OpenAI APIs, for image prompting generation, and put the key in ~/.mingdaoai/openai.key
  • Set up a key to use replicate.com APIs, and put the key in ~/.mingdaoai/replicate.key
• Create the outline of a presentation in a markdown file.
• Run the scripts to generate the presentation. They will
  • Step 1: convert the outline file into a file with title and image prompts
  • Step 2: generate images based on the image prompts
  • Step 3: combine the title and images into a presentation slide image
  • Step 4: combine all the slides into a final presentation

---

Cursor AI Prompt

You can save the following sections into a file called content.md, and let Cursor AI generate code based on the prompts.

Overview

The presentation system consists of the following scripts:

1. present1_content_generator.py - Generates slide content including titles, speaker scripts, and image prompts
2. present2_image_generator.py - Generates images for each slide based on the image prompts
3. present3_image_text_composer.py - Composes the final slides with images and text
4. present4_video_generator.py - Creates a video presentation from the composed slides

Required Files

Before running the scripts, ensure you have the following files in your presentation directory:

1. present.md - Contains the presentation outline and requirements
2. present_format.json - Contains the presentation format settings (dimensions, etc.)

Script Generation Instructions

1. Content Generator (present1_content_generator.py)

# Key components:
- SlideContent class for managing slide data
- SlideContentGenerator class for generating content using OpenAI API
- Support for both Chinese and English content
- Generates titles, speaker scripts, and image prompts
- Saves output to present_content.txt

2. Image Generator (present2_image_generator.py)

# Key components:
- SlideImageGenerator class for managing image generation
- Uses mutil.genPictureUtil for image generation
- Supports batch and single slide image generation
- Validates image dimensions and prompts
- Saves images in webp format

3. Image Text Composer (present3_image_text_composer.py)

# Key components:
- Composes final slides combining images and text
- Handles layout and positioning of elements
- Supports different presentation styles
- Generates high-quality slide images

4. Video Generator (present4_video_generator.py)

# Key components:
- Creates video presentation from composed slides
- Handles transitions and timing
- Supports background music
- Generates final video output

Usage Instructions

1. Create a present.md file with your presentation outline
2. Create a present_format.json file with your desired dimensions
3. Run the scripts in sequence: bash python present1_content_generator.py python present2_image_generator.py python present3_image_text_composer.py python present4_video_generator.py

Requirements

• Python 3.8+
• OpenAI API key (stored in ~/.mingdaoai/openai.key)
• Required Python packages:
• openai
• Pillow
• pathlib
• typing
• json
• argparse

Output Files

The scripts generate the following files:

1. present_content.txt - Contains all slide content
2. images/ directory - Contains generated images
3. present_*.webp - Individual slide images
4. present_*_prompt.json - Image generation prompts
5. present_*_prompt.txt - Text versions of prompts
6. Final video output

---

Ming Dao School uses 1-1 coaching and group events to help high-tech professionals grow their careers and handle career transitions.

If you like to join our upcoming mock system design interview events or other coaching programs, please contact us on LinkedIn.