Skip to content

bitandmortar/hyred

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Hyred (formerly local-resume-builder)

Overview

Hyred (Neural Resume Engine) is a local-first AI-powered application designed for tailoring resumes and cover letters with extreme privacy. It performs all processing on-device using a local LLM backend. The application matches a user's skills and prior experience against a job description, using RAG techniques to index background documents (like existing resumes or PDFs), ensuring honest and accurate, non-hallucinated tailored output.

Problem Addressed

Job seekers need to tailor their resumes and cover letters to stand out against Applicant Tracking Systems (ATS). However, sending highly personal work histories or resumes to online cloud LLM providers (e.g., OpenAI, Anthropic) poses a data privacy risk. Hyred addresses this by running offline on Apple Silicon, generating professional output with zero data leakage.

Approach & Capabilities

Hyred fuses a React/Vite-based frontend with a FastAPI backend.

  • Vision Parsing (M1 Vision): Extracts content from uploaded resume PDFs/images using Qwen2.5-VL-3B-Instruct hosted locally.
  • RAG & Tailoring (Ollama): Embeds and queries the user's extracted documents locally, feeding only factual work history alongside a job description into a local LLaMA 3.2 (3B) model.
  • ATS Optimization: Scores the generated resume against the target job description and highlights keyword match gaps.
  • Exporting: Results are exported as markdown files.

Installation & Setup Ensure you have Node.js and Python installed.

Backend Setup

  1. Validate Ollama is installed and running locally with llama3.2:3b.
  2. Validate M1 Vision endpoint is running (optional but required for PDF parsing).
  3. From the server/ directory:
pip install -r requirements.txt # (or install identical deps manually)
uvicorn server:app --host 0.0.0.0 --port 8020

Frontend Setup

  1. From the root directory:
npm install
npm run dev
  1. The UI will stream LLM responses and ATS scores locally at http://localhost:3015.

Usage

  • Drag and drop your baseline documents (PDF/docx/Markdown) onto the UI to upload.
  • Paste the description of the target role.
  • Specify your preferred LLM and the desired tone (e.g., formal or conversational).
  • Click Generate Neural Match to receive a tailored, ATS-ready markdown resume and a custom cover letter.

About

CV & Cover Letter Generator

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors