-
-
Ai movie maker interface
-
Home page
-
AI App Builder
-
Ai chat interface
Inspiration I have always been fascinated by the power of Artificial Intelligence, but I noticed a major friction point: users often have to jump between multiple platforms and tools to complete a single complex task. I envisioned a unified, cinematic "Command Center" that orchestrates the world's best AI models into one seamless workflow. My biggest inspiration was building something world-class from Sri Lanka, proving that even as a student, I can architect and deploy complex, multi-layered systems that solve real-world problems.
What it does PS JAVARS AI v0.2 is an advanced multi-tasking AI orchestration environment. It functions as a centralized dashboard that intelligently combines the strengths of Google Gemini, DeepSeek, and Claude. Its core capabilities include:
Cinematic Video Creation: An AI-driven engine to generate and assemble full-length video content.
Intelligent Problem Solving: Advanced mathematical reasoning and professional document synthesis.
Smart File Management: The ability to process, analyze, and categorize 100+ files simultaneously, including a specialized module for sorting audio by BPM.
Live Web Intelligence: Real-time web searching and deep-file analysis for data-driven insights.
How we built it The project was architected using the MeDo platform, which served as the backbone for our API orchestration.
MeDo Integration: We utilized MeDo to structure complex conversation flows and manage the back-end logic without the overhead of manual boilerplate code.
AI Orchestration: We integrated specialized APIs for specific roles—Google Gemini for visual/image processing, DeepSeek for logical reasoning, and Claude for creative linguistic tasks.
The Master Stitching Engine: A custom-built logic block designed to handle the rendering and merging of multiple AI-generated clips into a single output.
Significant Updates (Post-April 9) While the core vision of PS JAVARS began earlier, this project underwent a significant technical overhaul specifically for this hackathon period. Since April 9th, I have successfully:
Integrated Kling AI v1.5 as the primary high-fidelity image generation engine.
Developed the Real-time Progress Bar (%) system to provide visual feedback during heavy backend rendering tasks.
Re-engineered the BPM Audio Classifier logic to improve accuracy when processing batch music files.
Optimized the Multi-Agent workflow within MeDo to ensure seamless transitions between Gemini and DeepSeek models.
Challenges we ran into The primary challenge was balancing intense technical development with my O/L National Examinations. Managing study hours alongside debugging the Master Stitching Engine required extreme discipline. Technically, ensuring that the BPM Classifier could run consistently across different audio formats without latency was a significant hurdle that required multiple rounds of trial and error.
Accomplishments that we're proud of As a young developer, I am incredibly proud of building a system capable of managing 100+ files and rendering complex movie sequences. Successfully implementing a Progress Percentage system that accurately reflects backend processing was a major milestone, as it bridges the gap between a "prototype" and a "professional-grade" application.
What we learned This journey taught me the critical importance of System Architecture—how to facilitate communication between disparate AI models. I also learned that persistence is key; even under the stress of national exams, working on PS JAVARS kept my passion for innovation and engineering alive.
What's next for PS JAVARS AI v0.2 The roadmap for PS JAVARS is focused on autonomous collaboration:
Advanced Movie Maker 2.0: Enhancing narrative consistency across long-form video.
Batch Director Mode: Auto-stitching 20+ cinematic clips with seamless transitions.
3D Environment Integration: Moving beyond 2D video into interactive 3D virtual sets.
Voice-Command Integration: My ultimate dream is to evolve this into a fully autonomous assistant, making the science-fiction reality of J.A.R.V.I.S. accessible to everyone.
Legal & Submission Details Note to Judges: As I am currently a student under the age of 18, this project is submitted as a Team. My mother, M. D. Chathurika Priyamali, is officially designated as the Team Representative to ensure full compliance with the hackathon's legal and age requirements. However, I (Pahan Shaluka) am the sole creator, responsible for all conceptualization, system architecture, and technical development of PS JAVARS AI v0.2.
Log in or sign up for Devpost to join the conversation.