santifer/career-ops

An open-source job-search command center that turns any AI coding CLI into a pipeline to scan portals, score listings A-F, tailor CVs, and track applications locally.

What it is

career-ops runs inside an AI coding CLI (Claude Code, Codex, Gemini/Antigravity, OpenCode, Qwen, Grok, and others) and turns job hunting into a structured pipeline. It evaluates a job description against your CV, produces an A-F score, generates ATS-tailored PDFs, drafts outreach, and tracks everything in one place, while you keep the final decision. It is positioned as a filter for finding the few listings worth your time, explicitly not a mass auto-applier.

Key features

Auto-pipeline: paste a URL or JD and get an evaluation, a tailored PDF, and a tracker entry.
A six-block structured evaluation (role summary, CV match, level strategy, comp research, personalization, interview prep) plus a Block G posting-legitimacy check that flags scams and ghost jobs.
Portal scanner with 45+ preconfigured companies and 21 provider modules across Greenhouse, Ashby, Lever, and Wellfound; --verify runs a Playwright liveness check to drop expired postings.
ATS-optimized CV and cover-letter PDF generation from HTML templates via Playwright, plus draft-only application emails.
A Go/Bubble Tea terminal dashboard to browse, filter, and sort the pipeline.
Human-in-the-loop by design: the system drafts and recommends but never submits, sends, or clicks anything.
CLI-agnostic through the Agent Skill Standard, with a SKILL.md referenced per supported CLI.

Tech stack

JavaScript/Node.js, distributed as @santifer/career-ops (package version 1.22.0), run via npx.
playwright 1.61.1 for browser automation and PDF rendering; @google/generative-ai ^0.24.1, dotenv ^17.0.0, js-yaml ^4.1.1.
A dashboard TUI written in Go with Bubble Tea and Lipgloss (Catppuccin Mocha theme).
Data stored as Markdown tables, YAML config, and TSV batch files.
Scanner combines Playwright, the Greenhouse API, and WebSearch; batch processing uses headless CLI workers such as claude -p or opencode run.

When to reach for it

Running a focused, filter-first job search from an AI CLI you already use.
Generating per-listing ATS-tailored CVs and cover letters.
Tracking a pipeline of applications with automated merge, dedup, status normalization, and health checks.
Batch-evaluating many listings in parallel with headless workers.

When not to reach for it

You want spray-and-pray mass applying; the system recommends against applying to anything scoring below 4.0/5.
You expect fully automated submission; by design it never submits, sends, or clicks.
You have no AI coding CLI or no Node.js available.
You want a hosted, zero-setup service; this is a local tool where you supply your own model and keys.

Maturity signal

With roughly 61k stars, creation in April 2026, a push on the same day these facts were gathered, an MIT license at version 1.22.0, and around 200 open issues, this is a very young project growing and iterating rapidly. The large number of .mjs utilities and the high open-issue count point to fast, active development; the author notes the repo is maintained in about four hours a week with help from a fleet of AI agents.

Alternatives

feder-cr/Auto_Jobs_Applier_AI_Agent (AIHawk) — use it when you want full automated applying, which is the opposite of career-ops's human-in-the-loop, filter-first philosophy.
Reactive Resume — use it when you only need an open-source resume builder without portal scanning, evaluation, or tracking.

Notes

The core design principle is that career-ops is a filter, not an auto-applier: it drafts, scores, and ranks, but the human always submits. It can run entirely on free or local models (OpenRouter free models, Ollama, or any OpenAI-compatible endpoint), and even without installing a CLI via a standalone Gemini API script on the free tier. It is described as the first reference implementation of the CareerOps Manifesto, where signing the manifesto becomes a commit.

Wiki: santifer/career-ops