Hi everyone,
I'm Mekala Jaya Nanda Prabhas, a 3rd year Computer Science student at Siddhartha Academy of higher education. I've set up the App Inventor development environment locally, built the project successfully, and have been exploring the codebase
I'm interested in two GSoC 2026 project ideas:
Primary: Trainable ChatBot Interface and AI Component
I've studied how PIC (Personal Image Classifier) and PAC (Personal Audio Classifier) work as external training tools that export models for use inside App Inventor apps. I'd like to build a similar experience for chatbots — a web-based training interface where users can upload documents or FAQs as a knowledge base, and an App Inventor component that uses that knowledge base to respond to user queries at runtime.
I'm currently working on a prototype to explore how this could work end-to-end from the training interface to the App Inventor component integration. I plan to share it here once it's ready.
A few questions for the mentors (@natalie or whoever is mentoring this project):
- Should the training interface be built as a standalone web app (like PIC/PAC) or integrated into App Inventor's designer?
- For the runtime inference, is there a preference for on-device processing vs. a lightweight server-side approach?
- Are there any constraints on external dependencies or services that App Inventor projects should avoid?
Secondary: LMS Integrations (Google Classroom / Canvas)
I'm also interested in the LMS Integration project. I have experience with REST API integrations and OAuth flows, and I think connecting App Inventor with platforms teachers already use would be very impactful for the community.
I'm looking forward to contributing and learning from the community. Happy to hear any feedback or direction from the mentors.
Thanks,
Prabhas