Open tool automated 700+ job apps

A developer named Santiago has open-sourced a system called Career-Ops after using it on his own search for work. The repository says he used it to evaluate more than 740 job offers, generate more than 100 tailored CVs, and land a Head of Applied AI role. The project is built on Claude Code, Anthropic’s coding agent, and it turns a job hunt into a pipeline: scan listings, score fit, rewrite documents, track progress, and prepare for interviews (github.com, claude.com). What made the demo spread is the scale. Career-Ops says it can scan portals used by more than 45 companies, including employers such as Anthropic, OpenAI, ElevenLabs, Retool, and n8n, while also handling common recruiting platforms like Greenhouse, Ashby, Lever, and Wellfound (github.com). That is the part people notice first. One person, one terminal, hundreds of openings. The old spreadsheet version of a job search starts to look antique very quickly. But the repo is more revealing than the viral framing. It does not present itself as a bot that blindly fires applications into the void. The README says, in plain terms, that it is “NOT a spray-and-pray tool,” and the recent commit history shows the author adding ethical-use language and sharpening the message that the system should act as a filter, not a cannon (github.com, github.com). That distinction matters because the software is doing two different jobs at once. It automates the drudge work, and it also tries to decide which jobs are worth a human’s time. That second part is where the project becomes more than a clever macro. Career-Ops scores openings on 10 weighted dimensions, produces ATS-oriented PDFs tailored to a specific description, keeps a single application tracker, and stores material for behavioral interviews and salary negotiation (github.com). In other words, the system is not just filling forms. It is trying to build a consistent story about a candidate across the whole process. That is why the claim about 700-plus applications is a little misleading. The more interesting fact is that the tool was designed to narrow choices before submitting anything. Even so, the line between filtering and flooding is thin. A machine that can evaluate hundreds of roles in parallel can also tempt users to trust weak judgments at industrial scale. The repository itself hints at that risk. It tells users the first evaluations “won’t be great,” says the system needs onboarding like a new recruiter, and repeatedly insists on human review before submission (github.com, github.com). Those warnings are not decorative. They are admissions that the automation can move faster than the underlying judgment. The timing also helps explain why this landed so hard this week. Claude Code is having a moment of its own, with Anthropic marketing it as an agent that can work across a codebase, use command-line tools, and run parallel tasks from the terminal or an IDE (claude.com). Career-Ops shows what happens when that same agentic model leaves software engineering and walks straight into white-collar labor markets. The result is not a futuristic hiring platform. It is a job seeker building a private application factory, then publishing the blueprints on GitHub while the repo’s commit log fills up with fixes, translations, dashboard tweaks, and new ethical-use rules dated April 5 through April 7, 2026 (github.com).