Skip to content

joshbotz/applybrief

Repository files navigation

ApplyBrief

The brief you wish someone wrote for every job you applied to.

A couple of years ago I sent 350+ job applications over six months. A handful of interviews. Two offers. Two years later I still get rejection notices in my inbox. I built this for friends going through the same thing now, and I'm open-sourcing it so other people can use it too. It's a small command-line tool that finds jobs worth your time and writes the boring parts of the application for you. Honest, not hyped. Runs on your Mac or Windows PC. Free.

What you get

  • A short-list of jobs scored against your background. Not 500 listings — 10 to 30 that actually match what you said you want.
  • A one-page brief for any job you like. Where you fit, where you don't, a cover letter draft, interview prep, and a manual outreach plan for finding a warm intro on LinkedIn.
  • Honest about bad fits. If you're not a match, the brief tells you to skip — it won't pretend to write a cover letter for a job you shouldn't be applying to.

Install

Mac. Paste into Terminal:

curl -fsSL https://raw.githubusercontent.com/joshbotz/applybrief/main/install.sh | bash

Windows (10/11). Paste into PowerShell:

iex (iwr -useb https://raw.githubusercontent.com/joshbotz/applybrief/main/install.ps1).Content

Linux. Needs Python 3.11+ and pipx already installed:

pipx install git+https://github.com/joshbotz/applybrief.git@v0.1.0
What the installer actually does

It checks for a package manager (Homebrew on Mac, winget on Windows), Python 3.11+, and pipx — installs anything missing — then installs ApplyBrief itself from a pinned GitHub release tag. Mac details: INSTALL_NOTES.md. Windows details: INSTALL_NOTES_WINDOWS.md. No telemetry, no shell-config changes beyond pipx ensurepath.

Use it (3 commands)

applybrief init — one-time setup. Point it at your resume. Answer four short questions in your own words. Takes about a minute. Your profile lives on your machine.

applybrief discover — finds jobs. It searches LinkedIn, Indeed, Google Jobs, Glassdoor, and ZipRecruiter, then scores each posting 0–100 against your profile. You get a sortable HTML page that opens in your browser plus a markdown file.

applybrief apply <a job url> — go deep on one role. Give it a JD URL (or copy the URL and just run applybrief apply — it reads your clipboard). You get a one-page brief in your browser: fit assessment, resume tailoring notes, a draft cover letter (with a Copy button), interview prep, and a manual outreach plan.

You read each brief, edit the cover letter, and apply. If you can find a warm contact on LinkedIn, the brief tells you exactly who to look for — send a soft DM before you submit.

What it costs

Pennies a day with the recommended AI model. $0 if you run it locally with Ollama. You bring your own API key — there is no ApplyBrief subscription and no server.

Cost detail
Model Per Apply Brief Per discover run (~30 jobs)
Anthropic Claude Haiku (default) ~$0.001 ~$0.02
Anthropic Claude Sonnet ~$0.02 ~$0.50
Groq Llama 3.1 70B ~$0.0005 ~$0.01
Ollama (local) $0 $0

You pay the model provider directly. ApplyBrief takes no cut.

What it does with your data

  • Your resume and profile stay on your machine (~/Library/Application Support/applybrief/ on Mac, %LOCALAPPDATA%\applybrief\ on Windows).
  • They go to the AI provider you picked (Anthropic, OpenAI, Groq, or local Ollama) only when a brief needs to be written.
  • No tracking. No analytics. No server. The code is open source — read every line.

A few honest notes

  • This won't get you a job. You will. ApplyBrief saves you the hours you'd waste on the wrong applications.
  • Edit the cover letter. Recruiters can smell unedited AI output.
  • 5 great applications beats 50 mediocre ones. The whole point of the tool is to make "great" cheaper, not "mediocre" faster.
  • If you can find a warm intro on LinkedIn, send it before you submit. Referrals are 4–10x more likely to lead to a hire than cold applications.

What's not great yet (v0.1.0 is early — your feedback shapes v0.2)

I'm shipping this honest. Things that work well today:

  • apply produces high-quality briefs that refuse to fabricate (validated against 12 real cleared/federal JDs at avg 4.9/5 quality).
  • discover works great for defense / federal / cleared-roles space (Indeed indexes those densely).
  • HTML output is clean and shareable on both Mac and Windows.

Things that don't work well yet — please tell me when you hit one:

  • apply on LinkedIn / Indeed / Glassdoor URLs is unsupported (their pages are JavaScript-rendered and the fetcher can't read them). The workaround the tool tells you about: open the JD page, hit Cmd+A then Cmd+C, then run applybrief apply with no arguments — it picks up the JD text from your clipboard. Greenhouse, Lever, Ashby, and company career-page URLs work directly.
  • discover results vary by domain. Defense/cleared/federal: lots of dense, high-fit results. Cloud security, GRC, marketing, ops, sales: thinner pickings — many roles live on LinkedIn or niche boards that JobSpy can't easily scrape. Widening source coverage is on the roadmap.
  • Glassdoor and ZipRecruiter are often blocked without a proxy. They're in the default site list but frequently 400/403. LinkedIn, Indeed, and Google Jobs work reliably.
  • The fit scorer is sometimes generous with stretch roles. Read each brief — it's honest about gaps but the headline score can land 5–10 points high. Trust the gaps section, not just the number.
  • Windows install has been Mac-side dry-run validated but not yet tested on a real Windows box. If you're a Windows user trying this first — bless you, and please tell me anything that breaks.
  • No warm-intro engine yet. The brief tells you exactly who to look for on LinkedIn and includes a sample DM, but the lookup is manual. Building the engine is the next chunk of work.

Tell me what's broken or weak. Open an issue: github.com/joshbotz/applybrief/issues. Email me: josh@pillarsecurity.io. I'm building this with friends in mind — every report makes the next version better.

FAQ

Do I need a GitHub account? No. The install command above downloads what you need directly.

I've never used Terminal / PowerShell. Mac: Cmd+Space, type "Terminal", hit Enter. Windows: Start menu, type "PowerShell", click it. Paste the install command. Hit Enter. Type your password when it asks. That's the hardest part.

Is it really free? Yes. You pay the AI provider you choose (usually pennies a day) or run it for $0 with local Ollama.

Will this work on Windows? Yes — v0.1.0 ships with a native Windows installer.

What if I get stuck? Email me: josh@pillarsecurity.io. I read every one.

Roadmap

Warm-intro engine (LinkedIn 2nd-degree lookup) is the next big thing. Other stuff in CHANGELOG.md.

License & author

Apache 2.0. Built by Josh Botz — @joshbotz. If ApplyBrief helped you land something, I'd love to hear about it: josh@pillarsecurity.io.

About

Stop spraying 50 applications. Send 5 deeply-researched ones.

Topics

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors