Verified project record

ai-job-search-dev/ai-job-search

This project turns Claude Code into a job-application assistant that runs on your own machine, covering job-portal searching, tailored LaTeX CV and cover letter drafting, and interview preparation through slash commands. It requires a paid Claude subscription or Anthropic API credits plus a local toolchain including Python, Bun, and LaTeX.

Project overview

It packages the full job-search workflow—scraping, ranking, drafting, ATS-checking, and interview prep—into a single Claude Code-driven pipeline with a drafter-reviewer loop and PDF layout verification.

Project type: AI Agent · Workflow
Use cases: Documents & Office · Automation
Deployment: Refer to project documentation
License: MIT

Best for

General users who want to automate a full job search on their own machine and already have (or will set up) Claude Code with an active subscription or API credits.

Key capabilities

/scrape searches multiple job portals via CLI skills, deduplicates results, and presents them sorted by fit; /rank batch-scores postings into a ranked shortlist.
/apply runs a drafter-reviewer workflow: it evaluates fit, drafts a tailored LaTeX CV and cover letter, has a reviewer agent critique the drafts, revises them, then compiles and ATS-checks the PDFs. Every PDF is compiled and visually inspected, with targeted fixes applied until the layout is clean.
/interview builds a stage-specific prep pack from the application archive, maps likely questions to STAR examples, and offers a mock interview.

Limitations and risks

The job portal search skills are built for the Danish market; other markets need generated or copied skills. GPU and telemetry details are not documented. The data boundary is not documented end to end: local orchestration on the user's machine covers the workflow itself, while model inference runs through the paid Claude Code service, which requires an Anthropic subscription or API credits.

Getting started

Fork and clone the repo; install Python 3.10+, Bun, and a LaTeX distribution; run python3 tools/install_portal_clis.py; launch claude and run /setup; then run /scrape followed by /apply <url>. Setup difficulty is medium due to the required prerequisite toolchain.

Evidence and sources

GitHub project description: AI job application framework built on Claude Code: evaluate postings, tailor CVs, write cover letters, prep interviews.
README: I built this framework to run my own job search
README: A structured workflow that turns Claude Code into a full-stack job application assistant.
Release: v1.7.3 - Least privilege, postings checked before use, and fifty-eight fixes
README: Sixty-nine tailored applications, twenty first interviews, and one signed contract later

Source updated at: 2026-10-21T18:06:26Z

Open original GitHub project