Inspiration
Many students learn science through textbooks, diagrams, and memorized procedures because they do not have reliable access to laboratories, equipment, or individual guidance.
They may know the definition of refraction or surface tension, but they rarely get to practise the actual scientific process: making a prediction, performing an investigation, observing evidence, explaining the result, and defending a conclusion.
CurioLens AI was created to turn the objects already available around a student into meaningful, hands-on science learning.
What it does
CurioLens AI transforms everyday household objects into safe, guided science experiments.
A student can scan their surroundings or upload an image. CurioLens identifies available materials, matches them with suitable experiments from a human-reviewed safety library, and recommends activities that can be performed immediately.
The student then moves through a complete scientific learning cycle:
- Scan available objects
- Discover suitable experiments
- Follow safe, guided instructions
- Make a prediction before seeing the result
- Capture and analyse an observation
- Understand the scientific concept
- Complete a Claim–Evidence–Reasoning report
- Save the work to a persistent lab journal
CurioLens also includes a teacher workspace where educators can monitor class mastery, experiment engagement, student progress, and common misconceptions.
How it works
The student experience begins with object and material recognition. Detected materials are passed to an experiment recommendation engine, which compares them with the requirements of approved science activities.
For example, detecting a clear glass, water, and a spoon makes the Bending Light experiment available.
During the experiment, CurioLens provides:
- Safety guidance
- Step-by-step instructions
- Hints and simplified explanations
- Setup verification
- Prediction capture
- Observation interpretation
- Adaptive scientific explanations
- Socratic concept questions
- Claim–Evidence–Reasoning assessment
The completed activity generates skill scores for observation, claim, evidence, and reasoning. Results are saved in the student's lab journal and reflected in concept-mastery charts and achievement badges.
The teacher dashboard aggregates learning data to reveal class-wide patterns. Instead of only showing whether an activity was completed, it can highlight misconceptions such as students believing that a spoon physically bends in water.
Meaningful use of AI
CurioLens does not use AI as a decorative chatbot. Intelligence supports the learning journey through:
- Object and material recognition
- Experiment matching
- Setup verification
- Observation interpretation
- Adaptive explanation
- Student-response evaluation
- Misconception analysis
The application uses a modular AI service layer so recognition, recommendation, analysis, and evaluation providers can be changed independently. A local fallback keeps the core learning journey available when an external provider is unavailable.
Safety by design
CurioLens only recommends experiments from a controlled, human-reviewed library.
Experiments involving fire, electrical outlets, dangerous chemicals, sharp objects, pressurized containers, ingestion, medication, or body experimentation are excluded.
AI can personalize explanations, questions, and educational support, but it cannot freely invent unsafe experimental procedures.
How we built it
CurioLens AI was built with:
- Next.js and React for the application
- TypeScript for type-safe development
- Tailwind CSS for the responsive design system
- Zustand for persistent application state
- Recharts for progress and teacher analytics
- Framer Motion for interface transitions
- Browser camera and image-upload capabilities
- Speech recognition and text-to-speech where supported
- Local persistence for journals, mastery, settings, and active learning sessions
The application separates the interface, educational content, safety rules, state management, and AI services into independent layers.
Challenges we faced
One major challenge was designing a learning flow that used AI meaningfully without reducing the product to a chatbot.
Another challenge was managing experiment state across several stages: guided setup, prediction, observation, explanation, assessment, and reporting. We had to ensure that starting a new experiment cleared temporary progress while preserving completed journal history.
We also focused heavily on reliability. Camera permissions, unavailable external services, page refreshes, and repeated experiment attempts all needed graceful fallback behaviour.
Finally, fitting both a complete student experience and a useful teacher workspace into one intuitive product required careful navigation and interaction design.
Accomplishments
We are proud that CurioLens is more than a collection of educational screens. It supports a complete journey from ordinary materials to measurable scientific understanding.
Key accomplishments include:
- Material-aware experiment recommendations
- A complete guided Bending Light learning journey
- Prediction before observation
- Observation analysis and evidence identification
- Adaptive explanations and concept checks
- Claim–Evidence–Reasoning reports
- Persistent student journals and mastery tracking
- Teacher misconception and engagement insights
- A safety-controlled experiment library
- Responsive student and teacher interfaces
What we learned
We learned that educational AI is most valuable when it supports a clear learning method rather than simply generating answers.
Prediction, observation, explanation, and evidence-based reasoning create a much stronger learning experience than immediately revealing the correct concept.
We also learned the importance of separating temporary learning-session state from permanent student history, and of designing AI systems with safe and reliable fallback behaviour.
What's next
Future development will focus on:
- Expanding the approved experiment library
- Adding multilingual adaptive explanations
- Supporting authenticated student and teacher accounts
- Adding cloud synchronization across devices
- Enabling collaborative student investigations
- Packaging CurioLens as an installable offline-first application
CurioLens AI turns passive science content into active scientific discovery.
Your world is already a laboratory.
Built With
- api
- artificial
- computer
- css
- education
- hook
- intelligence
- learning
- machine
- motion
- next.js
- react
- recharts
- speech
- tailwind
- typescript
- vision
- web
Log in or sign up for Devpost to join the conversation.