Jobs_Applier_AI_Agent, called AIHawk in its README, is an MIT-licensed Python AI web agent that reads a resume, searches public job postings, scores the matches, and then applies to selected jobs through company application forms using a real Firefox browser, aimed at job seekers who want to automate repetitive applications while keeping the final choice.
What it is
Jobs_Applier_AI_Agent is a Python terminal application in the AI & Machine Learning / AI Development Platforms category. It combines resume parsing, local job search, match scoring and browser automation in one workflow: it reads a resume, finds jobs that fit from public sources, lets the user choose, and then applies through employer forms with a real Firefox browser.
The concrete problem it solves is the repetitive manual loop of searching job boards, opening postings and re-entering the same details into employer application forms. The README positions the project against LinkedIn Easy Apply and similar convenience flows: it requires no LinkedIn account and does not use Easy Apply, instead applying where the company takes applications. That makes it a replacement for the manual company-form application process rather than a job-board wrapper.
Key capabilities
- Reads the resume and turns it into a search covering the titles, skills and places that fit the applicant.
- Searches 10 million postings from 24 public sources, rebuilt every day, on the user's own computer.
- Reads the best 100 postings against the resume and scores the top 25 from 1 to 5, with the gaps and the reasons.
- Lets the user select the jobs they want and say yes once before applications are sent.
- Opens the company's form in a real Firefox session powered by invisible-playwright-mcp, fills every required field, attaches the CV and sends it, in a window the user can watch.
- Keeps what each page said, with a screenshot, for every application.
- Has no list of refused sites, and a site can be taught in one Markdown file, docs/sites.md, which records what has been confirmed.
Who uses it and how
- Individual job seekers on Windows or Linux, using the PowerShell run.ps1 installer or the shell run.sh installer for a terminal application.
- Applicants who want to apply through company forms rather than LinkedIn Easy Apply, because no LinkedIn account is required.
- Users who prefer local search, since 10 million postings rebuilt daily are searched on their computer rather than in a hosted service.
- People who want an audit trail, because every application retains a page record and a screenshot.
- Contributors extending coverage, using docs/sites.md for confirmed site support and docs/plugins.md for plugin-writing guidance.
Getting started
On Windows, install with the PowerShell command irm https://raw.githubusercontent.com/feder-cr/Jobs_Applier_AI_Agent/main/run.ps1 | iex; on Linux, use curl -LsSf https://raw.githubusercontent.com/feder-cr/Jobs_Applier_AI_Agent/main/run.sh | sh. Both routes require a key from OrcaRouter.
How it compares
The facts provide no list of paid products that this project replaces and name no direct competitors, so Jobs_Applier_AI_Agent stands alone in this registry. The one adjacent project named is invisible-playwright-mcp, which supplies the browser automation layer used by the agent rather than serving as a rival job-application product.
When to use it — and when not to
A self-hoster must run a local terminal install on Windows or Linux, keep a Firefox browser available and obtain an OrcaRouter key, because the README describes no hosted option. It is a poor fit for job seekers who want a fully managed service with no local installation or external key, or who need guaranteed coverage of every employer site without teaching any site definitions. The main limitation evident in the facts is that site coverage is extendable through a Markdown file and a confirmed-sites document, so the agent's reach depends on those definitions rather than a fixed universal adapter set.
