Recrux.AI — Intelligent Multi-Agent Job Matching

Status: Backend agents fully implemented. Frontend currently ships as a static UI showcase — backend wiring is the next milestone (see roadmap below).

A job-discovery and matching platform powered by a multi-agent AI architecture. Google Gemini + LangGraph orchestrate three specialized agents (resume parser, job searcher, match scorer) behind a FastAPI service. A separate React/Vite frontend shows the intended product UX.

Architecture

                    ┌─────────────────────────┐
                    │   React + Vite SPA      │
                    │   (frontend/)           │
                    │   static UI showcase    │
                    └───────────┬─────────────┘
                                │
                                │  (planned — currently mock data)
                                ▼
┌──────────────────────────────────────────────────────┐
│            FastAPI backend  (backend/)               │
│            CORS-restricted, env-driven origins       │
│                                                      │
│  /api/agents  →  LangGraph workflow                  │
│       │                                              │
│       ├──▶ resume_agent     (PyMuPDF / python-docx)  │
│       ├──▶ job_search_agent (mock JSON corpus)       │
│       └──▶ job_match_agent  (Gemini scoring)         │
└──────────────────────────────────────────────────────┘

Tech stack

Layer	Tech
Frontend	React 19, Vite 6, TypeScript 5.8, Tailwind 4, Lucide, Motion
Backend	FastAPI, LangGraph, LangChain, langchain-google-genai
LLM	Google Gemini
File parsing	PyMuPDF (PDF), python-docx (DOCX)
Data	Pydantic v2; mock job corpus in backend/utils/dummy_jobs.json

Run locally

Backend

cd backend
python -m venv venv
source venv/bin/activate            # Windows: venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env                 # then add your GEMINI_API_KEY
uvicorn main:app --reload            # http://localhost:8000

Open http://localhost:8000/docs for the interactive Swagger UI — the easiest way to demo the agents end-to-end without the frontend.

Frontend

cd frontend
npm install
cp .env.example .env.local           # then set VITE_API_URL
npm run dev                          # http://localhost:3000

Deploy

Get a Gemini API key first: https://aistudio.google.com/app/apikey.

Backend → Hugging Face Spaces (Docker, free)

1. Create a new Space at https://huggingface.co/new-space with SDK = Docker.
2. Push the backend/ directory to the Space (or link via the Spaces UI; HF auto-detects the backend/Dockerfile).
3. In Settings → Variables and secrets, add:
   • GEMINI_API_KEY (secret)
   • ALLOWED_ORIGINS = your Vercel frontend URL once it's deployed (e.g. https://recrux-ai.vercel.app)
4. The Space exposes port 7860; Swagger UI is available at /docs.

Frontend → Vercel (free)

1. Create a project at https://vercel.com/new, import this repo.
2. Root directory: frontend
3. Vercel auto-detects Vite. Build command: npm run build, Output: dist.
4. Environment variables: VITE_API_URL = your HF Space URL.
5. After first deploy, copy the Vercel URL back into the HF Space's ALLOWED_ORIGINS so CORS lets the frontend in.

Roadmap

• Wire frontend → backend. Add resume upload, dispatch to /api/agents, render scored matches.
• Live job APIs. Replace dummy_jobs.json with LinkedIn or JSearch.
• Vector store. ChromaDB integration is in requirements.txt but not yet wired into the agent graph.
• Auth. Google OAuth + email/password.

License

Reserved for Recrux.AI Team.