LitCompanion

Inspiration

As avid readers of English literature, which is not our native language, we frequently encounter complex plots, obscure metaphors, and challenging vocabulary. Switching between a book and a search engine or Goodreads to find context or recommendations is a constant friction point. We envisioned a reading buddy. This is not a distracting friend, but a supportive figure who explains deep concepts, translates phrases, and performs real-time research. Most importantly, we wanted this assistant to provide human-centric, empathetic responses inspired by Cognitive Behavioral Therapy (CBT) principles.

What it does

LitCompanion is an AI-powered reading assistant that bridges the gap between comprehension and emotional support. Through a voice-first interface (with a chat alternative), users can ask questions about their current read. By utilizing RAG (Retrieval-Augmented Generation) and personalized user surveys, LitCompanion provides answers that are not only factually accurate but also emotionally resonant and tuned to the user’s unique personality.

How we built it

Our core application is built with Next.js, Tailwind CSS, and Neon + Drizzle for rapid prototyping and robust data management. We use Google Auth for secure profiling, saving personality data to our Neon database. To bring the buddy to life, we integrated OpenAI for logic, Exa.ai for real-time web research, and ElevenLabs for expressive, diverse vocal performances. We specifically engineered prompts using CBT frameworks to ensure the AI's tone remains supportive and calming.

Challenges we ran into

Integrating Agora for real-time communication proved to be our steepest learning curve. We struggled with the architectural logic, particularly regarding speech interruption, which occasionally caused the system to hang. We also faced hurdles syncing ElevenLabs' dynamic audio with on-screen text highlighting. Since AI dictation is erratic compared to pre-recorded files, achieving perfect synchronization remains a work in progress.

Accomplishments that we're proud of

We are incredibly proud to have built a tool that promotes both literacy and mental well-being. Creating a seamless, clean interface that serves a meaningful purpose in such a short timeframe was our greatest achievement.

What we learned

This project was a masterclass in AI integration within the Next.js ecosystem. We moved from being intimidated by version control to mastering GitHub as a team, and we discovered the true power of collaborative problem-solving.

What's next for LitCompanion

We plan to scale for production and explore a subscription model to sustain the platform. Our ultimate goal is to make deep, comfortable reading accessible to everyone, regardless of their native language.