Note pilot is a study dashboard focused on helping users learn faster and more efficiently using LLMs in a more safer and reliable manner. Unlike regular chatbot sessions, lecture pdfs are used to generate study content based on those specific lectures such as flash-cards, exam style questions, summaries etc.
| Category | Tools |
|---|---|
| Framework |    |
| Language |  |
| Styling |  |
| Database |   |
| Deployment |  |
https://note-pilot-nu.vercel.app
- AI study content generation from uploaded lecture notes
- Secure authentication with dual token system (JWT + opaque token)
- Password reset with one-time cryptographic tokens via Mailjet
- E2E tested with Playwright against a containerized PostgreSQL.
- E2E: Playwright with containerized database via Docker
- Unit: Jest
Originally, the project utilized a simple HTTP-only cookie validated against the database on every request. To improve scalability and latency, the system uses a Hybrid JWT Model:
- Access Token: Short-lived JWT validated server-side for speed.
- Refresh Token: Long-lived token used to rotate access tokens, maintaining security without sacrificing user experience.
- User uploads lecture pdf: text data is extracted and stored in a relational database. This data is used as context when prompting the LLM, for the various features delivered.