AI Job Application Agent

Paste a job description, drop in your resume, and get a match score, tailored bullets, and a personalized outreach email or cover letter in seconds. Optional company research pulls live info about the company and weaves it into the output naturally.

Stack

Layer	Tech
Frontend	Next.js, TypeScript, Tailwind CSS
Backend	FastAPI, Python
AI	Claude (Anthropic)
Cache	Redis
Infra	Docker, Docker Compose
Hosting	Vercel (frontend), Railway (backend + Redis)

Features

• Match score — rates how well your resume fits the job (0-100) with a plain-English explanation of what fits and what's missing
• Tailored bullets — rewrites your master resume bullets to best match the job; never invents facts not in your original resume
• Email or cover letter — generates a concise cold outreach email or a formal cover letter, your choice
• Company research — live web search summary of what the company does, woven naturally into the output
• Auto-extract job info — detects the job title and company name from the pasted description automatically (can be toggled off to save API cost)
• Resume import — upload a PDF or DOCX and the text lands straight into the resume field
• Bring your own API key — paste your own Anthropic key to use your own quota; falls back to the shared key if left blank
• Job history — every run is saved in the browser with the score, output, tailored bullets, company research, and the full job description
• Redis caching — match scores, tailored bullets, job info extraction, and company research are all cached to avoid redundant API calls
• Rate limiting — 50 requests per IP per hour

Local Development

Prerequisites

• Docker Desktop
• Node.js 18+
• Python 3.11+
• Anthropic API key

Setup

1. Clone the repo

   git clone https://github.com/splotchysnow/ai-job-agent
   cd ai-job-agent

2. Configure the backend

   cd backend
   cp .env.example .env
   # Fill in ANTHROPIC_API_KEY in .env

3. Start the backend and Redis

   cd ..
   docker compose up --build

4. Start the frontend

   cd frontend
   npm install
   npm run dev

5. Open http://localhost:3000

Backend without Docker

cd backend
python -m venv venv
source venv/bin/activate   # Windows: venv\Scripts\activate
pip install -r requirements.txt
uvicorn main:app --reload

Redis is required for caching and rate limiting. You can run it locally with docker run -p 6379:6379 redis.

API Endpoints

Method	Endpoint	Description	Cached
GET	/health	Liveness check	-
POST	/match	Score resume fit against a job (0-100)	1hr
POST	/tailor	Rewrite resume bullets to match the job	1hr
POST	/draft	Generate outreach email or cover letter	-
POST	/extract-job-info	Extract job title and company from a job description	24hr
POST	/research	Research a company via live web search	24hr
POST	/extract	Extract plain text from a PDF or DOCX upload	-

Deployment

• Frontend — Vercel, auto-deploys on push to main
• Backend — Railway (railway up)
• Redis — Railway managed database

Environment Variables

backend/.env

ANTHROPIC_API_KEY=your_key_here
REDIS_URL=your_redis_url