Funil

Aplicación de gestión de notas con mínima fricción de usuario para proveer al usuario final de respuestas organizadas mediante procesamiento en segundo plano

Comment

Funil

Inspiration

We live surrounded by information.
Quick thoughts, saved links, songs we want to hear later, ideas we write down in a rush and forget the next day.

The problem isn’t capturing information. We’re actually very good at that.
The problem is finding it again when we need it.

Funil was created to solve that exact frustration.

Instead of being just another place to dump notes, Funil is designed to filter and organize what you save, so your ideas don’t get lost in the noise.

What it does

Funil is a web app where you can:

  • Write text notes
  • Save links (YouTube, Spotify, or any URL)
  • Upload files and images
  • Combine different types of content into a single entry

But the key difference is what happens behind the scenes.

When you create a note:

  • If you don’t give it a title, Funil creates one for you.
  • The content gets structured automatically.
  • It understands what kind of resource you added.
  • It can extract useful details like dates or intentions.
  • It can even detect if something sounds like a future reminder.

Funil doesn’t just store information.
It helps make it usable.

How we built it

Backend

  • Java 21
  • Spring Boot
  • REST API architecture
  • JPA / Hibernate
  • Relational database
  • Local LLM integration using Ollama

Frontend

  • React
  • TypeScript
  • Minimalist, clean design focused on clarity

How it works

  1. The user creates a note with different resources.
  2. The backend gathers the content and prepares context.
  3. A structured prompt is sent to a local language model.
  4. The model returns enriched content (like a generated title or structured output).
  5. The processed data is saved in the database.

The goal was simple: keep the system structured, while letting AI enhance the experience without taking control away from the user.

Challenges we ran into

  • Designing prompts that produce consistent results.
  • Deciding how much responsibility to give the AI.
  • Handling different types of content in a single entry.
  • Keeping the architecture clean and scalable.
  • Avoiding unnecessary complexity.

The biggest challenge was finding the right balance between automation and control.

Accomplishments that we're proud of

  • Fully working local LLM integration.
  • Reliable automatic title generation.
  • Clean, modular architecture.
  • A product concept aligned with its visual identity.
  • A system that feels simple to use but powerful underneath.

Most importantly, we built something that solves a real, everyday problem.

What we learned

  • AI is a tool, not the product itself.
  • Good structure makes intelligence more effective.
  • Simplicity requires discipline.
  • Clear architecture makes future improvements easier.
  • Building something useful is better than building something flashy.

What's next for Funil

This is just the beginning.

Next steps include:

  • Semantic search across notes
  • Automatic tagging and categorization
  • Smart reminders
  • Native integrations with Spotify and YouTube
  • Personal insight dashboard
  • Mobile optimization

Our long-term vision is simple:

Funil becomes a reliable, structured extension of your memory.

Share this project:

Updates