Inspiration
We noticed the difficulties in discovering the aesthetic and vibe of a book solely based on its cover and synopsis. This leads to many books going unnoticed and readers struggling to find books they would enjoy reading.
Currently, short-form platforms such as TikTok are trending for their ability to engage viewers with content. BookTok aims to bridge the gap between books and short-form entertainment by using AI to generate animations based on a book’s synopsis, reviews, and author input. Through creating short-form content that is easily accessible to readers, BookTok helps readers discover new stories and allow authors to overcome the challenges of being discovered.
What It Does
BookTok has many different features including:
- Allowing authors to upload their book summary and keywords that describe the vibe
- AI generated animation that reflects the synopsis of the book
- A scroll tab where readers can scroll through different animations to find a book that suits them
- A library to save liked books so readers can refer to future reads
- A search bar where users can type in book names and AI uses external sources to generate an animation based on summaries, themes, and reviews
- A digital bookshelf where books users have read can be stored as well users can import their Goodreads so books they have already read are included in their digital library
- Users can set a reading goal and when they have reached their goal they can unlock different features to add to their bookshelf like book nooks
- Authors can also view the analytics to see where readers interest dropped off
- Users can like and react to specific moments in the story with emojis
- Authors also have the option to externally contact animators and upload those animations instead of the AI generated one
How We Built It
We built BookTok as a mobile app using React Native and Expo with a TikTok-style scroll-based video feed. The backend uses Node.js, Express.js, and MongoDB Atlas (with Mongoose), deployed on Render.com. When a book is added:
- Google Gemini 2.0 Flash analyzes the summary, extracts tropes, and generates visual prompts.
- Google Imagen 3 creates atmospheric book-themed images from those prompts
- ElevenLabs generates a voiceover using the "Rachel" voice model
- Fluent-FFmpeg compiles the images and audio into a short video
- Cloudinary stores and serves the generated videos.
- The final video URL is saved to MongoDB and displayed in the app's feed.
Accomplishments
- Using AI text analysis to create visual animations which help represent the aesthetic of a book
- An app that combines short-form entertainment, a reading log, and boosts author discoverability
- Creating a gamified digital bookshelf which rewards consistent reading
- Providing analytics to authors so they can receive feedback on their books
What We Learned
We learned the power of AI in storytelling and entertainment. We gained insight into how powerful of a tool AI can be when converting keywords into full animations which provide more visual detail than the book cover. We also better understand the value of keeping our customer engaged which is why we added features such as the Goodreads import to help minimize manual processes allowing for better user engagement.
Challenges
- Turning abstract keywords into visual concepts
- Managing external sources and ensuring they are reliable and safe platforms
What’s Next for BookTok
We would like to expand:
- Style customization of the animations, make the animations not only through keywords but also through stylistic words to help improve the accuracy of the vibe
- Add a community tab where people can connect to book clubs online or in their area and also find other people with similar book taste
- Develop the scroll tab to use an AI that can better capture what the user likes, what animations they linger on longer, and which animations they scroll past to better devlope the AI to recommend animations that are better-suited for the reader
Built With
- elevenlabs
- expogo
- express.js
- ffmpeg
- gemini
- javascript
- mongodb
- mongodbatlas
- node.js
- react-native
Log in or sign up for Devpost to join the conversation.