Inspiration
- Many students, and people in general, want to learn new skills but getting started can feel overwhelming. One of the hardest part is knowing where to begin
- Most real-world skills don't come with a clear roadmap and without structure, learners may feel lost, lose motivation, or abandon learning the skill entirely
- Since many people thrive in structured, step-by-step learning, a roadmap to guide you through learning could help them
What it does
RoadRunner helps users break down any skill they want to learn into a structured, actionable learning path
- AI-Generated Learning Roadmaps
- Generates personalized roadmaps based on a user's learning goal
- Organizes skills into
- Curates learning resources for each step
- Dependency Checking
- Identifies prerequisite skills before advancing
- Ensures the roadmap follows a valid learning order
- Prevents gaps in foundational knowledge
- Feedback and Knowledge Checks
- Quizzes users on topics to mark them as completed or not
- Provides AI-generated feedback based on responses
- Encourages mastery of prerequisites before progressing further
How we built it
- Frontend
- React + Electron for the core desktop application
- TailwindCSS for styling
- Javascript for fast iteration
- Backend
- OpenRouter and Gemini API for agentic reasoning, grading, and structured AI outputs
- Drizzle ORM for database maangement and response caching
Challenges we ran into
- Resolving merge conflicts while working in isolated environments
- Maintaining consistent data flow between services
- Refactoring and debugging styles from vanilla CSS to TailwindCSS
- Learning how to use Electron
Accomplishments that we're proud of
- Coordination as a team of 4
- Built working end-to-end application that generated a detailed roadmap for a skill
- Developed a system that integrated AI services and APIs
- Adaptability to picking up Electron and making APIs
What we learned
- Hands-on experience with AI platforms such as OpenRouter and Gemini
- Building Electron applications and handling interprocess communication across system
- System prompt designing AI agents for structured output handling and agentic reasoning
What's next for Road Runner
- A dynamic roadmap that adapts each time the user returns by asking what they’ve recently learned, automatically updating completed tasks and expanding the roadmap as they progress
- Voice input support for knowledge checks to make assessments more accessible and natural
- Video-based skill verification, allowing users to upload recordings instead of traditional quizzes to validate hands-on or physical skills
Built With
- drizzle
- electron
- gemini
- javascript
- openrouter
- react
- tailwind
- vite

Log in or sign up for Devpost to join the conversation.