(this program is a failure yet)
I executed the project end-to-end as a solo contributor. I learned Chrome Manifest V3 and Chrome’s built-in AI sessions from scratch, implemented the six prompt “skills” (Prompt/Writer/Rewriter/Proofreader/Translator/Summarizer) behind a session manager, added a deterministic fallback for when on-device AI isn’t available, and built the privacy dashboard and DevTools panel to keep behavior transparent. I wrote the prompts, guardrails, and UX, handled local caching and integrations (GitHub/StackOverflow/MDN), and produced the docs, demo assets, and evaluation guide. The hardest part was climbing the learning curve while keeping the user experience fast, safe, and explainable.
Bullet list (résumé style)
Sole creator: architecture, implementation, prompts, and UX
Learned Chrome MV3 and on-device AI APIs during the build
Implemented six prompt skills + language detection with guardrails
Built privacy dashboard, DevTools panel, and deterministic fallback
Wrote docs, demo flow, and evaluation guide for judges
Log in or sign up for Devpost to join the conversation.