HackStack

Project Info

Ballad AI

Devpost

This project did not submit a demo video on Devpost.

Inspiration

Reading is dying in 2025. With attention spans shortened by TikTok and Netflix, books feel static and boring compared to multimedia entertainment. As avid readers ourselves, we wished for a method to make book more immersive. We asked: what if every story had its own dynamic soundtrack that adapted to the narrative?

What it does

Ballad AI analyzes any book using natural language processing to understand emotional beats, pacing, and atmosphere. It then generates personalized musical accompaniments that sync with the reading experience—building tension during climactic scenes, adding mystery to detective stories, or creating wonder during fantasy adventures. Readers get an immersive, multimedia experience that makes books better.

How we built it

We combined several AI technologies: Text Analysis: Language models to extract emotional sentiment, mood, and narrative pacing from literary text. Music Generation: Generates dynamic soundscapes based on textual analysis Real-time Sync: Built a reading app that reads the book aloud and adjusts music in real-time Web Interface: Created a responsive web application using React for seamless user experience

Challenges we ran into

Fine-grained Music Control: We originally envisioned layering many music tracks so that we can dynamically shift the mood of the music in real time. However, current available APIs only offer text-conditioned generation with a maximum of 30s, so we had to compensate with a fading and looping effect. Processing Time: Current music generation is slow, so it's difficult to generate music in real time.

Accomplishments we're proud of

Successfully generated coherent, emotionally-appropriate soundtracks for diverse genres Created smooth transitions between musical themes that enhance rather than distract from reading Built a working prototype that demonstrates the concept's viability

What we learned

Our frontend was partially developed by AI, but we found AI generated code to be difficult to maintain. Most of our members had to pick at least 1 unknown framework such as FastAPI, Next JS, etc.

What's next

Implementing user voice recognition so we can allow for the user to read the book aloud Eye tracking, generating visual elements to accompany the music and narrative Better musical control with Lyria Realtime API

Analysis

Compare with all teams

View

Metric

Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.

Technology

Found in codeClaimed only
  • CSSIn code
  • FastAPIIn code
  • Next.jsIn code
  • PythonIn code
  • ReactIn code
  • Tailwind CSSIn code
  • TypeScriptIn code
  • AnthropicClaimed

7 of 8 appear in the indexed code. 1 claimed on Devpost could not be matched to code, which may simply mean the tool leaves no trace in the repository.

AI coding agents

No AI coding agent signals were found in this repository.

Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.

Codebase size

Source size

359 KB

Source files

73

Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.

0 stars