study-ops

Open-source AI study search: scan university portals, evaluate listings into a structured A–H report with a global 1–5 score, tailor your CV and statement of purpose, track applications — runs locally in your AI coding CLI.

Reverse-engineered from career-ops, an AI job-search pipeline, retargeted from hiring to admissions. SPEC.md records what was ported, what changed, and why.

Status: MVP

• Evaluate a program, PhD position, or scholarship from a URL or pasted text

→ A–H report + tracker row

• Deadlines — what's due, sorted, with referee ask-by dates
• CV — render and tailor from cv.md
• SOP — context assembly, word-limit enforcement, fabrication scan
• Verify — flag claims that appear in no primary source file

Not built yet (spec'd in SPEC.md Part IX and Part XII): the portal scanner, batch processing, document intake, interview prep, outcome analytics.

Quick start

It tells you exactly what is missing. Then, in your CLI:

What makes it different from a job-search tool

Four things are not renaming exercises:

1. Deadlines are fixed and annual. Missing one costs a year. Job postings are rolling; you apply when ready. So the tracker has a Deadline column, an expired deadline stops an evaluation before it starts, and deadlines.mjs computes the referee ask-by date — because referees are the long pole and the only part of an application you don't control.