AI-Powered Short-Form Video Automation System

We are looking for an experienced AI automation developer to build a complete system that automatically creates and publishes short-form videos for YouTube Shorts, Instagram Reels, and TikTok. The system should take a topic or keyword as input and automatically generate: A video script Voiceover using TTS A complete vertical short-form video Captions or subtitles Title, description, hashtags, and platform-specific post content Automatic publishing to YouTube, Instagram, and TikTok Important Requirement The final output must be an actual video with moving footage, animations, generated video clips, or other dynamic visual content. We are not looking for a simple image slideshow with voiceover. AI and LLM Requirements Because API costs can become expensive, we prefer using a local LLM and locally hosted AI tools whenever practical. However, we are also open to paid AI APIs when they provide significantly better video quality or reliability. The system should be designed so that: The LLM provider can be changed easily Local and cloud-based AI models can be switched through configuration Different TTS or video-generation services can be added later The entire system can be installed easily on another computer or server API keys, models, languages, and output settings can be changed without modifying the core code A Docker-based setup or another portable installation method is preferred. Core Features Topic Input The user enters a topic, keyword, or short description. Script Generation The AI creates an engaging short-form video script optimized for platforms such as YouTube Shorts, Instagram Reels, and TikTok. Language Selection The user must be able to select the output language. Initial languages may include English and Korean, but the system should support adding more languages later. Text-to-Speech The system generates natural-sounding voice narration based on the selected language. Local TTS solutions are preferred when the quality is acceptable. Video Creation The system automatically creates a vertical video, preferably in 9:16 format. The video should include: Dynamic footage or generated video clips Voiceover synchronization Automatic subtitles Background music or sound effects Transitions and basic motion graphics Configurable video duration and style Again, the video should not be only a collection of static images. Review and Approval Ideally, the user should be able to preview the generated script and video before publishing. It should also be possible to regenerate or manually edit the script, voice, title, description, or video. Automatic Publishing The completed video should be published automatically to: YouTube Shorts Instagram Reels TikTok The system should generate platform-specific titles, captions, descriptions, and hashtags. It should also support scheduled posting if possible. Technical Preferences We are open to your recommended technology stack, but the system should be: Modular Maintainable Easy to install Easy to move to another computer or server Easy to switch between local models and paid APIs Well documented Designed with error handling and logging Built with secure API credential management Potential technologies may include Python, Node.js, Docker, FFmpeg, ComfyUI, Ollama, local TTS models, video-generation APIs, and official social media APIs. You do not need to use all of these tools. Please recommend the most reliable and cost-effective architecture. Deliverables Fully working source code Installation and setup documentation Docker or simple deployment configuration Configuration file for models and API providers Automated short-form video generation workflow YouTube, Instagram, and TikTok publishing integration Instructions for installing the system on another computer Documentation explaining how to replace the LLM, TTS provider, or video-generation provider Basic testing and bug fixes after delivery When Applying Please include: Examples of similar AI video automation or content-generation systems you have built Your recommended local LLM, TTS, and video-generation tools Which parts can realistically run locally Which parts may require a paid API and the estimated usage costs How you would generate actual dynamic video rather than an image slideshow Your recommended hardware requirements Your estimated development timeline Your fixed-price estimate or milestone breakdown Please do not send a generic proposal. Briefly explain the architecture you would recommend for this project.