Code Gatekeeper

Code Gatekeeper is a conversational AI platform that transforms how developers learn and retain understanding of AI-generated code—addressing the critical “AI Paradox” where increased code production leads to decreased comprehension and critical thinking skills.

---

Core Functionality & Working Prototype

Our fully functional prototype features two sophisticated ElevenLabs Conversational Agents:

1. Voice Quiz Agent

Acts as a “code gatekeeper” that verbally quizzes developers on AI-generated code before allowing approval.
It asks probing questions about:

• Logic
• Edge cases
• Implementation details

2. Voice Tutor Agent

Provides personalized, topic-specific voice lectures on code concepts, adapting explanations based on the submitted code context.

---

Technical Complexity & Multimodal Orchestration

• ElevenLabs Conversational AI — Real‑time voice‑to‑voice interaction with dynamic prompt injection
• Claude AI (Anthropic) — Intelligent code analysis, question generation, and educational content creation
• Clerk Authentication — Secure user management and session handling
• React + TypeScript — Modern, type‑safe frontend architecture
• Dynamic Prompt Engineering — Runtime context injection for personalized agent behavior

---

Innovation & Creativity

Unlike traditional code review tools, Code Gatekeeper uses voice-based Socratic dialogue to ensure developers genuinely understand code before shipping.

This groundbreaking approach:

• Forces active recall through verbal explanation
• Creates accountability in AI-assisted development
• Preserves human critical thinking skills in the age of AI

---

Real-World Impact

Addressing a $50B+ problem in enterprise software: technical debt from poorly understood code.

Scalable to:

• Engineering teams requiring code comprehension gates
• Educational institutions teaching programming
• Bootcamps ensuring genuine skill development
• Enterprise compliance environments requiring code understanding audits

---

Theme Alignment

Code Gatekeeper exemplifies conversational agents by unifying voice, browser, and AI APIs into a cohesive autonomous assistant that listens, reasons, responds, and evolves—perfectly embodying the hackathon mission to:

“Build something that talks back.”

---

None