Revolutionizing Content Creation: A Glimpse into My Automation Project.

Over the past month, I embarked on an exciting project that merges my passion for technology with creative content creation led by OpenAI's custom chat GPT.

This journey has been about exploring the possibilities of automation in generating short movies, publishing scenery images, and managing social media content. Here's a glimpse into my project's core components and achievements:

(1): Developed custom chat GPT ??: only 5 steps for setting up all necessary files and data for short movie generation.

(2): Streamlined Short Movie Production ??: Achieved approximately 50% automation in creating engaging short films and published 50 videos or more publicly.

(3): Automated Publishing ??: Achieved approximately 80% automation in publishing scenery images and shared over 100 scenic images with zero manual uploads.

Launching on Social Media: A Milestone

The launch of @quickquips60 across Instagram, TikTok, and YouTube shared over 50 videos and marked a significant milestone in my project.

• TikTok: https://www.tiktok.com/@quickquips60
• Instagram: https://www.instagram.com/quickquips60/
• YouTube: https://www.youtube.com/@quickquips60

Also, over 100 scenery images have already been published without manual uploads or typing underscores the potential of automation in creative content generation.

• Instagram: https://www.instagram.com/view_vistas/

Diving Deeper into My Automation Journey:

1. Custom Chat GPTs: A Core Innovation

I've taken Chat GPTs to the next level, tailoring them specifically for my project's needs. After three significant updates in the last 30 days, my system now not only checks custom instructions against a knowledge file but also retrieves the instructions from the GitHub repository through API calls, ensuring outputs adhere closely to pre-defined formats. This has drastically reduced setup times for content, making it possible to prepare everything needed for a short movie in under five minutes. For example, in one of the operations, as follows, the GPT will execute the story script creations and DALL.E image prompts. Both formats are provided in the code snippet, so I can swiftly move to the next actions.

As well, to save time for image generation, I instruct the GPT to create a single four-panel strip image, rather than four images individually. The other purpose of this approach is to ensure a character design and image style as consistent as possible; however, sometimes the GPT shows different character designs on each panel but generally maintains consistency overall. The following examples show the images with higher consistency (Image 01) and partial inconsistency (Image 02).

2. Short Movie Creation: 50% of Automation

Tools:

• Python libraries (notably OpenAI, cv2, moviepy, etc.)
• Custom chat GPT (introduced above)
• Buffer, a software application designed for managing multiple accounts for social networks.

Utilizing those tools, I've managed to automate half of the short movie creation process. The reason for 50% of automation is because of mixing manual operations and automation:

Manual:

• Ask the GPT to show unique punchlines, story scripts/DALL.E image prompts(demonstrated above), landscape images for the title and last page, and messages for the SNS account.
• Set the schedule posts on Buffer, an SNS management service, which allows users to publish videos on multiple social media accounts at the same time.

Automation:

• Crop the four-panel strip image into four scenes and add some decorations on each.
• Generating video and audio based on story scripts with background music.

To be clear, once all the data and files (story scripts, four-panel strip image, and images for title/last pages), the video will be created with 100% automatically. Every timing and animation is predefined; that was the hardest work for me due to complicated math calculations. Please look at an example video and how I succeeded in setting up the transition timing on each scene and the volume of background music.

3. Image Publication: Achieving 80% Automation

Tools:

• Python Codes
• Custom chat GPT
• Google Drive API
• Zapier, the provider for creating the automatic workflow by integrating multiple APIs from various third parties.

By integrating the tools above, I've streamlined the process of publishing high-resolution scenery images. This sophisticated workflow automates the generation of image descriptions and hashtags, uploads to Google Drive, and posts to social media through Zapier, all without manual intervention.

Here are the detailed processes:

1. Python codes connect to the Google Drive API and upload the image file in PNG format with a description; this description contains short explanations and hashtags.
2. Zapier workflow detects the updates of the new image files.
3. Zapier connects to the Instagram account, and then publishes the uploaded images with a caption; this caption will be retrieved from the "description" metadata of the image file.

Finally, only a single update of the image publishes the content automatically. Leveraging this automatic system, over 100 images have already been published successfully without any manual typing and posting.

Thank you for reading this article in the end. The next personal project has already been launched; probably the financial modeling for the stock index.

Stay Tuned!!!