Instaclip: AI Short-Form Video Content Generator

Instaclip leverages cutting-edge AI technology, including OpenAI's Whisper for automatic transcription and GPT-3 for content analysis, alongside MoviePy for video editing, to transform long videos into captivating short-form content.

Designed for creators aiming to enhance their presence on platforms like Instagram Reels and TikTok, Instaclip automates the selection and editing process, making content creation effortless and accessible.

Features

Whisper Integration: Utilizes OpenAI's Whisper to accurately transcribe audio, enabling the AI to understand and select key moments based on both video and audio content.
GPT-3 Content Analysis: Employs GPT-3 to analyze transcriptions and identify the most engaging segments, ensuring the clips resonate with audiences.
MoviePy Editing: Leverages MoviePy for efficient video editing, providing high-quality, customizable short-form videos.
Customizable Lengths: Supports generation of videos tailored to the specific requirements of different social media platforms.
Batch Processing: Offers the ability to process multiple videos simultaneously, streamlining content creation.

Installation

Prerequisites

Python 3.8 or higher
FFmpeg
An OpenAI API key for Whisper and GPT-3

Steps

Clone the Instaclip repository:

git clone https://github.com/jabezborja/instaclip.git

Navigate to the Instaclip directory:

cd instaclip

Install the required dependencies:

pip install -r requirements.txt

Usage

Before running Instaclip, ensure you have set your OpenAI API key in your .env:

OPENAI_API_KEY='your_api_key_here'

To convert a long video into short-form content, follow these steps:

Launch Instaclip:

flask --app main run

Note: The frontend is still implementing so you can use the API endpoint with POSTMAN for now.

To generate a short-form content. You must call

/video/upload -> uploads the videos and turns it into listenable mp3 for Whisper
/video/segmentation -> you pass the video_filepath from /video/upload, this returns segmented and processed transcription
/video/segment_candidates -> you pass the processed segmented and transcription, this returns the video candidates - best possible content
/video/export -> you pass the video_filepath and candidates, this returns the output path of the videos, this is now the short-form content generated.

Contributing

We welcome contributions to Instaclip. To contribute, please follow these guidelines:

Fork the repository.
Create a new branch for your feature (git checkout -b feature/AmazingFeature).
Commit your changes (git commit -m 'Add some AmazingFeature').
Push to the branch (git push origin feature/AmazingFeature).
Open a pull request.

License

Instaclip is licensed under the GNU Affero General Public License (AGPL). This means that if you use the software (or any part of it) to interact with users over a network, you must provide the source code under this license.

See the LICENSE file in the repository for the full license text.

sidaga/instaclip