Application Tracking System

AI-powered resume analysis and cover-letter generation built with Next.js, Stack Auth, Convex, and Google Gemini.

Features

• Resume analysis with match scoring, strengths, weaknesses, skills match, and recommendations
• Cover-letter generation with configurable tone and length
• Tailored LaTeX resume generation using selectable templates
• PDF resume upload and parsing
• Searchable history for analyses and cover letters
• Auth-protected dashboard flows
• Rate limiting (Upstash Redis in production, in-memory fallback for local/test only)

Tech Stack

• Next.js 16 (App Router)
• React 19 + TypeScript 5
• Tailwind CSS 4
• Convex (data storage and queries)
• Stack Auth
• Google Gemini (@google/generative-ai)
• Radix UI + Lucide icons
• Vitest + Testing Library

Getting Started

Prerequisites

• Node.js 18+
• Bun 1.0+
• A Google Gemini API key
• A Convex project URL
• Stack Auth project credentials

Installation

1. Clone and install dependencies

git clone https://github.com/Aditya190803/application-tracking-system.git
cd application-tracking-system
bun install --frozen-lockfile

1. Create local env file

cp .env.example .env.local

1. Set required variables in .env.local

# Gemini
GOOGLE_API_KEY=""

# Convex
NEXT_PUBLIC_CONVEX_URL=""

# Stack Auth
NEXT_PUBLIC_STACK_PROJECT_ID=""
NEXT_PUBLIC_STACK_PUBLISHABLE_CLIENT_KEY=""
STACK_SECRET_SERVER_KEY=""

# Optional
NEXT_PUBLIC_APP_URL="http://localhost:3000"
MODEL_NAME="gemini-2.5-flash"
AI_TIMEOUT_MS="30000"
PDF_PARSE_TIMEOUT_MS="12000"
COVER_LETTER_ROUTE_TIMEOUT_MS="35000"
RESUME_ROUTE_TIMEOUT_MS="45000"
LATEX_RENDER_API_BASE="https://latexonline.cc"
UPSTASH_REDIS_REST_URL=""
UPSTASH_REDIS_REST_TOKEN=""
# Only set true for local emergency fallback; keep false/empty in production
ALLOW_IN_MEMORY_RATE_LIMIT=""

# Axiom (production logging)
AXIOM_TOKEN=""
AXIOM_DATASET=""
AXIOM_SERVICE="ats"
LOG_OBS_ERROR_DETAILS=""

1. Start Convex dev backend and Next.js app

bunx convex dev
bun run dev

App runs at http://localhost:3000.

Observability (Axiom)

Server routes emit structured JSON logs ([obs] prefix) via src/lib/observability.ts.

1. Create an Axiom dataset (e.g. ats-prod).
2. Create an API token with ingest permission.
3. Set on Vercel (and in .env.local for dev ingest):
4. AXIOM_TOKEN — bearer token (never commit)
5. AXIOM_DATASET — e.g. ats-prod
6. AXIOM_EDGE — optional, e.g. us-east-1.aws.edge.axiom.co (matches your ingest URL region)
7. Optional: add the Vercel → Axiom integration for raw platform logs in addition to direct ingest.

Useful Axiom queries

['your-dataset']
| where event == "pdf.parse_failed" or event == "pdf.parse_no_text"
| sort by _time desc

['your-dataset']
| where requestId == "paste-ref-from-user"

PDF upload failures include requestId in the API error; the upload UI shows it as (ref: …) so users can share it with support.

Set LOG_OBS_ERROR_DETAILS=true temporarily to include sanitized errorMessage on failure events (never resume or job description content).