Customize your LaTeX resume and cover letter for every job application, automatically. Paste a job description, get a tailored and optimized one-page resume and matching cover letter as compiled PDFs.

You write one master resume in LaTeX with all your experience, projects, and skills. The tool reads a job description and uses an LLM to tailor your resume at every level:

• Sections (Education, Experience, Projects, Research) are included or excluded based on relevance.
• Items (individual jobs, projects, publications) are selected or dropped per role.
• Bullets within each item are individually included or excluded, with minor keyword edits to match the JD language.
• Skills are reordered by relevance, with new skills added and irrelevant ones removed per category.
• Cover letter is generated from scratch using your selected resume content and the JD.
• One-page enforcement automatically drops the lowest-scored optional content and recompiles until it fits.

The output is a compiled PDF ready to submit. Every run is archived with timestamps so you never lose a previous version.

git clone https://github.com/avishj/AutoCustomizeResume.git
cd AutoCustomizeResume
uv sync

Then follow the setup instructions in examples/ to configure your resume, API key, and config file.

• Python 3.10+
• uv (recommended)
• Tectonic on PATH
  • macOS: brew install tectonic
  • Linux: install guide
  • Cargo: cargo install tectonic
• LLM API key for any OpenAI-compatible provider (or local Ollama)

You mark sections, items, and bullets with LaTeX comment tags so the tool knows what it can include or exclude.

%%% BEGIN:pinned:education        ← always included
\section{Education}
    %%% BEGIN:optional:state-u     ← LLM decides per section, per item, per bullet, and then can edit bullet too!
    \resumeSubheading{State University}{...}
    %%% END:optional:state-u
%%% END:pinned:education

%%% SKILLS:languages               ← LLM can reorder, add, remove
\textbf{Languages}{: Python, Go, Rust, SQL}
%%% END:SKILLS:languages

Full specification: docs/TAGS.md

Start from examples/config.example.yaml and copy it to config.yaml.

Naming template variables: {first}, {last}, {company}, {role}, {date}, {timestamp}

Process a single job description and exit.

uv run autocustomizeresume --jd path/to/jd.txt

Monitors jd.txt (configurable via paths.jd_file) and rebuilds on every save.

uv run autocustomizeresume

Press Ctrl+C to stop.

uv run autocustomizeresume --jd jd.txt --company "Acme Corp" --role "Backend Engineer"

Any OpenAI-compatible API works. Set llm.base_url and llm.model in config.yaml, and put the API key in .env.

"tectonic is not installed or not on PATH" Install Tectonic and make sure it is accessible from your terminal. See Prerequisites.

"API key not found" Set the environment variable named by llm.api_key_env (default: LLM_API_KEY) in your .env file or shell.

"Config file not found" Copy examples/config.example.yaml to config.yaml in the project root.

Resume exceeds one page after all retries The tool auto-drops lowest-scored optional items, but if your pinned content alone exceeds one page, reduce content or adjust template spacing. You can also convert some pinned tags to optional to give the tool more flexibility.

Watch mode not triggering Check that you are editing the file at paths.jd_file (default: jd.txt). The watcher debounces changes by watcher.debounce_seconds before running.

GNU Affero General Public License v3.0 (AGPL-3.0)