Chronos AI

Inspiration

History is usually taught as isolated facts, dates, and static articles. We wanted to build a platform where history can be explored as a connected knowledge graph—allowing users to navigate civilizations, people, wars, empires, and events visually while asking natural language questions.

Our vision is to create an AI-powered historian that helps users understand not only what happened, but also why it happened through evidence-backed historical data.

What it does

Chronos AI transforms structured historical datasets into an interactive historical knowledge platform.

Current features include:

Importing historical entities and events from Wikidata
Enriching historical records with Wikipedia summaries and metadata
Storing historical knowledge in Amazon Aurora PostgreSQL
REST APIs for:
- Historical entities
- Events
- Relationships
- Search
- Scenarios
Relationship graph generation
Timeline-ready event retrieval
Map-ready geographic data
Foundation for an AI Historian retrieval system

Our long-term vision includes:

Interactive historical timelines
Dynamic historical maps
Relationship graph exploration
Evidence-backed AI conversations
Alternate history simulations

How we built it

The application is built using a modern full-stack architecture.

Frontend

Next.js 15
React
TypeScript

Backend

Next.js Route Handlers
TypeScript
REST APIs

Data Pipeline

Historical information is ingested from Wikidata using SPARQL queries. The imported entities and events are then enriched using the Wikipedia REST API before being stored in Amazon Aurora PostgreSQL.

The backend exposes APIs for entities, events, relationships, search, scenarios, timelines, and maps, allowing the frontend to consume structured historical knowledge.

Challenges we ran into

Building a historical knowledge platform presented several challenges:

Designing a schema flexible enough to represent civilizations, people, battles, wars, and empires.
Building an idempotent ingestion pipeline that can safely rerun without duplicating data.
Mapping multiple historical data sources into a consistent model.
Securely connecting Amazon Aurora PostgreSQL during development.
Preserving source attribution for future evidence-backed AI responses.

Accomplishments that we're proud of

Built a complete historical data ingestion pipeline.
Connected Wikidata and Wikipedia into a unified historical dataset.
Designed a scalable historical knowledge graph model.
Built REST APIs powering entities, events, relationships, search, and scenarios.
Established the backend foundation for an AI Historian capable of retrieval-augmented historical reasoning.

What we learned

This project reinforced that history is naturally represented as a graph rather than isolated records.

We also learned that trustworthy AI applications require structured retrieval, source attribution, and clean historical data before introducing large language models.