Inspiration

Students today have access to endless information, but having information available does not mean they will remember it when it matters. Many students spend hours studying yet still struggle because they forget key concepts before exams.

We wanted to build an AI learning companion that moves beyond summarising notes. Future Rescue was inspired by the idea that AI should not only answer questions after a student is stuck, but proactively identify what they are likely to forget and help them reinforce it before it disappears.

What it does

Future Rescue is an AI-powered learning companion that predicts what students are likely to forget and creates personalised rescue missions to strengthen their memory.

Students provide their study notes, and Future Rescue:

• Extracts the key concepts from learning material
• Builds an understanding of the topics being studied
• Predicts which concepts are most likely to be forgotten
• Explains why those concepts are at higher risk
• Generates personalised rescue missions using learning techniques such as active recall and spaced repetition

Instead of simply telling students what they studied, Future Rescue helps them understand what they need to remember next.

How we built it

Future Rescue was built using:

• Next.js with App Router
• TypeScript
• Tailwind CSS
• shadcn/ui components
• OpenAI models for the AI intelligence layer

The application uses AI-powered workflows for:

1. Concept extraction — turning unstructured study notes into structured learning concepts.
2. Forgetting prediction — analysing concept difficulty, complexity, and relationships to estimate memory risk.
3. Rescue mission generation — creating personalised learning actions designed to reinforce weak areas.

Codex was used throughout development to accelerate implementation, create reusable components, debug issues, and iterate on the product architecture.

Challenges we ran into

One of the biggest challenges was deciding what information was actually useful for a student. A simple notes summariser was not enough — we needed to create something that could reason about learning and memory.

Another challenge was designing AI outputs that felt trustworthy. Instead of making absolute predictions, we focused on explaining why a concept may be difficult and providing actionable recommendations.

Building a complete AI workflow while keeping the experience simple and intuitive was also a major design challenge.

Accomplishments that we're proud of

We are proud that Future Rescue demonstrates a complete AI learning loop:

Study notes → Concept understanding → Forgetting prediction → Personalised intervention

The most important achievement is that the AI does not stop at analysis. It takes action by generating rescue missions that help students improve their retention.

We also built a complete demo experience that allows users to see the value of the product within seconds.

What we learned

We learned that effective AI products are not just about generating responses — they are about understanding the user's goal and creating meaningful next steps.

Throughout development, we learned how to combine AI reasoning with thoughtful product design, build reliable structured outputs, and use AI as both a development tool and the foundation of the product itself.

What's next for Future Rescue, The Forgetfulness Abolishing Learning Helper

Future Rescue has many opportunities to grow.

Next steps include:

• Supporting more learning formats such as PDFs and lecture materials
• Creating long-term memory tracking for students
• Adapting predictions based on individual learning behaviour
• Adding calendar integration for timed rescue missions
• Building richer progress insights over time

The long-term vision is for Future Rescue to become a proactive AI learning companion that helps every student remember what matters, exactly when they need it.

Built With

• gpt-5.6
• next.js
• tailwind