Type any job title → get a complete, visual career roadmap. Skills, certifications, who's hiring, salary expectations, free courses & YouTube channels — plus an AI advisor you can chat with.
Built at VibeThon 2026.
🔗 Live demo: https://pathforge-ai-studio.vercel.app
You tell PathForge the job you want — e.g. "Data Scientist" — and it instantly builds:
| Feature | What you get | |
|---|---|---|
| 🗺️ | Visual roadmap | The stages from beginner → hired, each with how long it takes and what to learn |
| 🛠️ | Skills needed | The exact skills employers look for |
| 🏅 | Certifications | Real certs worth getting, with links |
| 💰 | Salary expectations | Entry, mid, and senior pay ranges |
| 🏢 | Who's hiring | Companies with open roles for that job right now |
| 📺 | Learn from the best | Free courses + top YouTube channels |
| 🤖 | AI advisor | Chat and ask anything ("Do I need a degree?") and get an honest, grounded answer |
Everything is generated live from the web + AI — nothing is hard-coded.
You (browser)
→ Frontend (React website, hosted on Vercel)
→ Backend API (FastAPI / Python)
→ Google Gemini → writes the roadmap, skills, salaries
→ Tavily + Firecrawl → search the live web for courses & YouTube
→ Remotive + Arbeitnow (free job boards) → who's hiring
→ AI Advisor → n8n workflow → Gemini → answers your questions
The frontend asks the backend one question ("give me a roadmap for X"). The backend talks to several free services at the same time, blends the results, and sends back one tidy package the website draws on screen.
| Folder | What it is |
|---|---|
frontend/ |
The website users see (React + TanStack Start + Tailwind + shadcn/ui) |
backend/ |
The brain (FastAPI). Turns a job title into the full roadmap. Includes backend/n8n/ — the chatbot workflow |
prompts/ |
Every AI prompt we wrote, documented. (Judges: this is our prompt engineering) |
agent-specs/ |
Our build plan — we split the work across 4 AI coding agents; these are their briefs + the shared API contract |
lovable-prompts/ |
The prompts my teammate used to build the frontend in Lovable |
cd backend
python -m venv .venv
# Windows:
.venv\Scripts\activate
# macOS/Linux:
# source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env # then fill in your free keys (see below)
uvicorn main:app --port 8000Check it: open http://localhost:8000/api/health → you should see {"status":"ok"}.
No keys yet? Set
MOCK=1in.envand the backend serves realistic sample data offline — great for a quick look.
cd frontend
npm install
cp .env.example .env # VITE_API_URL=http://localhost:8000
npm run devOpen the URL it prints (usually http://localhost:5173), type a job title, and hit Forge My Path 🚀.
Put these in backend/.env (copy from backend/.env.example):
| Key | Where to get it (free) | Used for |
|---|---|---|
GEMINI_API_KEY |
https://aistudio.google.com/apikey | Roadmap, salaries, chat (model: gemini-2.5-flash) |
TAVILY_API_KEY |
https://tavily.com | Web search for courses, YouTube, salary sources |
FIRECRAWL_API_KEY |
https://firecrawl.dev | Web scraping fallback |
N8N_WEBHOOK_URL |
your n8n (import backend/n8n/careerpath_rag.json) |
The AI advisor chatbot |
Companies-hiring uses Remotive and Arbeitnow, which need no key.
- If any data source is slow or down, that section degrades gracefully — the app never crashes.
- Two Gemini keys with automatic failover when one hits its rate limit.
- Mock mode (
MOCK=1) for a guaranteed offline demo.
Frontend: React 19 · TanStack Start · Tailwind · shadcn/ui · deployed on Vercel Backend: FastAPI (Python) · Google Gemini 2.5 Flash · Tavily · Firecrawl · Remotive + Arbeitnow Chatbot: n8n (Webhook → AI Agent → Gemini)
Built at VibeThon 2026.
- Backend, Integration & Deployment — Rudranil Chatterjee
- Frontend, Design & Interactivity Testing — Manika Agrawal Gogoi
Proprietary — All Rights Reserved. See LICENSE. This project may be viewed for evaluation/judging only. No copying, modification, distribution, or use of any part (code, prompts, or design) is permitted without the Authors' prior written permission.
Made with way too much coffee. ☕