ApplyLint

ApplyLint is an AI workflow that prepares, audits, and refines job application cover letters before submission.

Instead of generating unverified text, ApplyLint runs an adversarial review loop: it evaluates candidate fit, drafts an initial letter, checks it against 32 canonical recruiter rejection codes, and rewrites the content to fix grounding errors without inventing missing qualifications.

Demo

ApplyLint-Demo-Walkthrough.mp4

---

Resume + Job Description
           ↓
     Match Analysis
      ↙          ↘
[not_a_fit]    [fit]
    ↓             ↓
   END      Cover Letter Generation
                  ↓
            Recruiter Lint (Attack)
                  ↓
       Cover Letter Improvement (Defense)
                  ↓
                 END

---

Core Principles

1. Candidate Data is the source of truth: The model cannot invent skills, metrics, job titles, or dates.
2. Transferable experience boundary: Adjacent experience explains adaptability, but is never presented as existing experience with an unfamiliar tool.
3. Adversarial evaluation: Recruiter Lint flags claims unsupported by Candidate Data or in conflict with timeline constraints.
4. Candidate in control: ApplyLint prepares, analyzes, and improves applications. It never submits them on the user's behalf.

---

Technical Documentation

Detailed documentation for each stage of the pipeline:

• AI Workflow Overview: Complete multi-stage LangGraph pipeline and streaming state lifecycle.
• Match Analysis: Candidate-job requirement matching, scoring caps, and fit recommendations.
• Cover Letter Generation: Initial generation modes and grounding constraints.
• Recruiter Lint: 32 canonical reason codes, severity levels, and adversarial validation contracts.
• Cover Letter Improvement: Strategy matrix for resolving findings truthfully.

---

Tech Stack

• AI & Orchestration: LangGraph, LangChain, Ollama
• Backend: FastAPI, Python 3.10+, SQLAlchemy, aiosqlite, Alembic
• Frontend: Angular 22, Tailwind CSS, Spartan UI
• Streaming: Server-Sent Events (SSE) for live node updates

---

Getting Started

Prerequisites

• Python 3.10+
• Node.js (tested on v24.19.0) and npm
• Ollama running locally

1. Backend Setup

cd backend
python -m venv .venv
source .venv/bin/activate

pip install -r requirements.txt
alembic upgrade head
uvicorn main:app

The API runs at http://localhost:8000.

2. Frontend Setup

cd frontend
npm install
npm run build
npm run start

The frontend runs at http://localhost:4200.

---

Origin

ApplyLint started as a 48-hour build-in-public marathon project to address hallucinations and ungrounded claims in AI job applications. It is under active development to add resume tailoring, automated file ingestion, and a dedicated verification stage.