About

Built Because ATS Screeners Are Broken

Every existing ATS checker uses the same generic algorithm and charges you for it. None of them simulate how real enterprise platforms actually parse your resume. So I built one that does. For free.

The Problem

Why Existing ATS Screeners Fail You

Generic Algorithms

They all use the same basic keyword-matching approach regardless of which ATS your application will actually go through. A score from "ATS Checker Pro" means nothing if the company uses Workday.

Paywalls Everywhere

Most "free" ATS checkers give you a teaser score then lock detailed results behind a paywall. Job seekers already under financial stress shouldn't have to pay to optimize their resume.

Data Privacy Concerns

You upload your resume (personal info, work history, contact details) to a random website. Where does that data go? Who has access? ATS Screener parses your files entirely client-side. Only extracted text is sent to Google Gemini for AI scoring.

No Real ATS Knowledge

Workday parses resumes completely differently from Greenhouse. Taleo's boolean keyword filters work nothing like Lever's context matching. A one-size-fits-all score is useless.

The Approach

How ATS Screener Works Differently

01

LLM-Powered Analysis

Instead of hard-coded keyword matching, ATS Screener uses large language models to semantically analyze your resume against each ATS system's known behavior. The LLM understands context, transferable skills, and industry-specific terminology across any field.

02

Six Real ATS Profiles

ATS Screener simulates Workday, Taleo (Oracle), iCIMS, Greenhouse, Lever, and SAP SuccessFactors. Each profile reflects the parsing strictness, keyword strategy, and scoring behavior of that platform based on research into publicly available documentation.

03

Built-In Fallback

The primary scoring engine uses Google Gemini AI. If the AI provider is unavailable, a deterministic rule-based engine automatically provides scoring as a fallback. You always get results.

04

Client-Side First

Resume parsing (PDF/DOCX text extraction, section detection, entity extraction) happens entirely in your browser via Web Workers. Only the extracted text is sent to Google Gemini for scoring. Your resume files never touch any server.

Under the Hood

Tech Stack

Frontend

  • SvelteKit 5 (Svelte 5 runes)
  • Scoped CSS + CSS custom properties
  • Motion.dev for animations
  • Bits UI for accessible primitives

Engine

  • Custom PDF/DOCX parser (pdfjs-dist + mammoth)
  • TF-IDF + NLP tokenizer
  • 6 ATS profile scorers
  • Skills taxonomy (500+ terms, 10+ industries)

AI

  • Gemma 3 27B via Google (primary)
  • Llama 3.3 70B via Groq (fallback)
  • Rule-based fallback engine
  • TF-IDF keyword matching
  • Skills taxonomy (500+ terms)

Infrastructure

  • Vercel (hosting)
  • Firebase Auth + Firestore
  • GitHub Actions (CI/CD)
  • Vitest + Playwright (testing)

100% Open Source

Every line of code is public. No premium tiers, no paywalls. MIT licensed. Fork it, improve it, self-host it.

Creator

Built by Sunny Patel

Software engineer who got tired of paying $30/month for ATS tools that give the same useless generic score. Built this as an open-source alternative that actually simulates real enterprise platforms.