MOIRAI

MOIRAI breathes life into your novels. Plot your character's emotional journey, create unique visualisations and get suggestions for finishing your book.

Comment

Inspiration

Every creative knows how hard it is to finish a project. As a fiction writer, I wanted an AI tool to help me craft narrative storylines for my characters, inspired by the existing literary corpus. I wanted to query books not by their author or resume, but by their storylines.

What it does

MOIRAI is:

  • a database of characters' emotional arc
  • a GUI to visualise characters and their arc
  • A tool to automatically extract characters and storylines from a user-provided novel

How I built it

I began by creating a database of characters and their arc using a third-party Python library called BookNLP. Once I had a working prototype running on the Railway hosting platform, I deployed it to Google Cloud. I then created a new version of the database by using Gemini agents.

I use Gemini's coding capabilities through and through to help me navigate the APIs, the tech stacks i was not familiar with.

Challenges I ran into

Learning cloud deployment. Running out of Gemini credits :')

Accomplishments that I am proud of

I'm proud of:

  • the original idea
  • climbing the learning curve

What I learned

Everything.

  • Database: learned SQLite, PostgreSQL, Gemini's gen-ai API, NLP techniques, prompt engineering
  • BackEnd: learned FastAPI, sqlalchemy, pydantic
  • FrontEnd: learned to use React, Vite and npm.
  • FullStack: learned to coordinate volumes, PostgreSQL database, front and back ends.
  • DevOps: learned to use Docker, testing environments
  • Hosting: learned to use Google Cloud CLI

What's next for MOIRAI

Continuing MOIRAI, I will:

  • Process more books to complete the database
  • Build a business-facing book recommender system that includes the characters' storylines
  • Leverage Gemini's API and connect it to the database to build a client-facing tool to craft storylines.
Share this project:

Updates