Lintrix

Lintrix uses Gemini 3 to teach coding through the Socratic method. Our AI asks guiding questions instead of giving answers, helping you think like a real programmer. lintrix.codes

Comment

Inspiration

Learning to code is hard. But what's harder? Learning when every AI tool gives you the answer instead of teaching you how to think. We watched countless students copy solutions from ChatGPT, pass their exercises, and still fail technical interviews. The problem isn't that AI is too smart—it's that it's too helpful. We wanted to create an AI that teaches like a great professor: by asking the right questions at the right time.

What it does

Lintrix is an AI-powered coding education platform with a Socratic tutor. Instead of providing direct answers, our Gemini 3-powered AI asks guiding questions that lead students to discover solutions themselves. Features include: an interactive code editor, real-time code execution, AI-generated algorithm flowcharts, unlimited personalized exercise generation, deep-dive concept explanations, and a gamified progress system with points and achievements.

How we built it

Frontend: React 18 + Vite + TailwindCSS with Monaco Editor for the IDE experience Backend: FastAPI (Python) deployed on Google Cloud Run AI: Gemini 3 Flash API for all 5 AI features (tutor, flowcharts, exercises, explanations, code analysis) Code Execution: Piston API for secure sandboxed code running Auth & Database: Firebase Authentication + Firestore Hosting: Firebase Hosting with custom domain (lintrix.codes)

Challenges we ran into

Making AI NOT give answers: The hardest part was prompt engineering Gemini to respond with questions instead of solutions. We iterated dozens of times to perfect the Socratic behavior. Mermaid flowchart syntax: AI-generated diagrams sometimes had syntax errors. We implemented validation and fallback mechanisms. Rate limiting: Balancing API quota while keeping the experience smooth required careful backend proxy design. Real-time feedback: Synchronizing code execution results with AI analysis without lag was technically challenging.

Accomplishments that we're proud of

Successfully implemented 4 distinct Gemini 3 integrations in a single application Created an AI that genuinely teaches instead of just answering Built a complete, production-ready platform deployed at lintrix.codes Achieved sub-2-second AI response times Designed a beautiful, intuitive UI that makes learning feel like a game

What we learned

Prompt engineering is an art—small wording changes dramatically affect AI behavior The Socratic method is powerful but requires careful implementation to avoid frustrating users Gemini 3's speed makes real-time educational AI applications truly viable Building for education requires balancing guidance with challenge

What's next for Lintrix

Interview Simulation Mode: Practice mock coding interviews with AI interviewers Multiplayer Challenges: Compete with friends in real-time coding battles More Languages: Expand beyond Python to JavaScript, Java, C++ Mobile App: Native iOS/Android experience Enterprise Version: For coding bootcamps and universities AI Code Review: Detailed feedback on code quality, not just correctness

Built With

  • css
  • fastapi
  • firebasefirestore
  • firebasehosting
  • googlecoudrun
  • googlegemini3
  • html
  • httpx
  • javascript
  • lucide
  • mermaidjs
  • monacoeditor
  • pistonapi
  • pydantic
  • python
  • react
  • tailwindcss
  • vite
Share this project:

Updates