How to use ai-job-search: setup, commands and tips

What you need before you start

The project requires Claude Code, which the README describes as having no free tier: you need a Claude Pro, Max or Team subscription, or Anthropic API credits. You also need Python 3.10 or newer, Bun, and a LaTeX distribution that includes lualatex and xelatex.

• Claude Code (CLI), with a paid Claude plan or API credits
• Python 3.10+ for the tools in tools/
• Bun, which runs the job search CLI tools
• A LaTeX distribution with lualatex and xelatex (TeX Live, MacTeX, TinyTeX or MiKTeX)
• pypdf (optional), installed with pip install pypdf, for the ATS parseability check in /apply

Fork and clone the repository

Run the fork-and-clone command from the README, then move into the new folder. Be aware that a fork of a public repository is always public on GitHub, and /setup writes personal data such as name, contact details, employment history and salary expectations into tracked files. If the copy is for your own search rather than for contributing back, the README recommends a private repository with this repo as upstream instead; the recipe is in SETUP.md section 8.

Install the job search tools

Run the installer script from the repository root. On Windows the README gives the variant that uses python instead of python3. The installer runs bun install in each portal CLI under .agents/skills/*/cli/, skips any CLI with install-time scripts and reports it, and supports a --dry-run flag that shows what would be installed.

• linkedin-search and freehire-search have zero runtime dependencies and run with plain bun

Set up your profile with /setup

Open Claude Code in the project folder and run /setup. It offers three paths: read your documents/ folder (CV PDF, LinkedIn export, diplomas, reference letters, past applications), import a single CV pasted in chat, or walk through an interview. It detects what you have and asks. The documents-folder mode can be re-run safely as you add material; documents/README.md describes the layout.

Search and rank job postings

/scrape searches multiple job portals for positions matching your profile, deduplicates results and presents them sorted by fit. If a scrape returns more jobs than you want to review by hand, /rank batch-scores them against the fit framework and gives you a ranked shortlist first. The portal search skills are built for Danish boards such as Jobindex, Jobnet and Akademikernes Jobbank, and the README says the pattern can be swapped for your local job boards.

Apply to a job

Pass a job posting URL to /apply. The command drafts a tailored CV and cover letter. The README notes that if a URL cannot be fetched, because some portals block automated access, you will need another route into the posting. Output CVs are compiled with LaTeX, and the cover letter uses xelatex. Compiling the CV with pdflatex often fails on modern MiKTeX installs, so use lualatex as the README specifies.

• CV compiles with lualatex
• Cover letter compiles with xelatex because cover.cls requires fontspec
• Optional pypdf check for ATS parseability