TruthLens

BCSAI · Hackathon Project · Automate Learning

Turn messy notes into verified knowledge.

TruthLens is a note-taking workspace that fact-checks your study notes with AI — inspecting every sentence, flagging weak or incorrect claims, and handing back corrections with reasoning and sources. All without breaking your flow.

my-study-notes.md

The mitochondria is the powerhouse of the cell. Water boils at 90 °C at sea level. The Great Wall of China is visible from space with the naked eye. Python is a dynamically typed language.

4 sentences · not yet analysed

01 — What it is

A personal fact-checker, built into your notebook.

Most learners take notes quickly without verifying them, miss subtle factual errors, and spend hours double-checking. TruthLens automates that feedback loop so notes become reliable knowledge.

🔬

Sentence-level intelligence

Inspects every sentence you write and flags weak or incorrect knowledge.

🤖

AI-powered fact checking

Uses local LLMs via Ollama to verify claims and ground corrections.

🔗

Source citations

Each correction comes with references to trustworthy sources.

📝

Effortless corrections

Get refined notes with confidence that every claim is verified.

📊

Learning workspace

A dashboard to manage documents, view analysis, and track corrections.

🔒

Secure & private

User authentication with secure, per-user workspace management.

02 — How it works

Six steps from raw notes to reliable knowledge.

Create an account

Sign up and get your personal workspace.

Write your notes

Create documents and start writing study notes.

AI analysis

Run TruthLens to analyse each sentence for factual accuracy.

View results

See flagged sentences with corrections, reasoning, and sources.

Apply corrections

Review and apply suggested fixes to improve your notes.

Build knowledge

Maintain a reliable, fact-checked knowledge base over time.

03 — Architecture

How the pieces fit.

A Next.js front end talks to a Django REST backend, which persists to PostgreSQL and calls a local Ollama LLM engine for the actual fact-checking.

Next.js
frontend
Django
REST backend
PostgreSQL
database
Ollama
LLM engine

04 — Stack

Built with.

Frontend

  • Next.js 16
  • React 19.2
  • TypeScript
  • Tailwind CSS
  • Framer Motion
  • shadcn/ui · Radix

Backend

  • Django 5.0
  • Django REST
  • PostgreSQL 15
  • Django Unfold
  • CORS Headers

AI / ML

  • Ollama (gpt-oss:20b)
  • PyTorch
  • HF Transformers
  • Prompt-engineered fact checking

DevOps

  • Docker
  • Docker Compose
  • Multi-container

05 — What inspired it

Built for students like us.

"We both go to an international university where professors' first language is often not English — and it becomes incredibly hard to take notes and make sure what you wrote is right." — the TruthLens team
PRIMARY Automate Learning: Build Smarter Study Tools
📚 GitBook Best Documentation
🤖 Built with Cline CLI