hire_ex_¶
Need a job? Sip your drink.
HireEx reads the boards for you.
Here's the idea. You tell HireEx a bit about yourself once — what kind of job you want, what you don't want, what you'd never do. Then you forget about it. Every morning it's read through a few hundred postings for you and handed you the handful that are actually worth your time.
Nothing leaves your laptop except the actual job descriptions, which go to the AI model you pick. No account. No tracking. No monthly bill.
Mostly, you just sip your drink and click apply.
Start here¶
One command, three minutes
Openshttp://localhost:3000. Paste an OpenRouter key in Settings and you're triaging.
- Getting started — install it, run it, set up your profile.
- Privacy and data — what stays local, what goes where.
- Configuration — every setting, field by field. For the curious.
- Architecture — developer notes on how it's wired.
A typical morning¶
| Time | What happens |
|---|---|
| 08:00 | 482 fresh postings pulled across 13 boards |
| 08:03 | 174 obvious no's dropped by a cheap AI pass |
| 08:04 | 63 duplicates skipped (same job on two boards) |
| 08:07 | 245 ranked — 3 worth your attention today |
| 08:07 | You sip your coffee and click apply |
How it thinks¶
Six steps, all on your laptop. A seventh — the background hunter — is optional and catches companies the boards haven't seen yet.
Pulls fresh postings from thirteen boards at once: LinkedIn, Indeed, Glassdoor, Google Jobs, ZipRecruiter, Bayt, Naukri, Greenhouse, Lever, Ashby, Workable, RemoteOK, WeWorkRemotely. If something crashes mid-way, it picks up where it left off.
Companies you've blacklisted, titles you've said no to — gone on the spot. Edit the list any time from the dashboard; past postings from that company disappear too.
A cheap AI model does a fast read of every posting and drops obvious no's. Five buckets, firm rules. If it's unsure, it rejects — better safe than sorry.
Same job on LinkedIn and Indeed? HireEx notices and only shows you one. This part runs entirely on your machine — no calls out to the cloud.
A stronger AI model reads the survivors against your profile and scores each one: how well it fits, how the tech matches, salary, growth, risks. Writes a short explanation, even hints what to tweak in your resume for this role.
The dashboard. Glance at scores, keyboard-shortcut through them, archive the chaff. Five minutes a morning instead of two hours.
=== "Background hunter" (optional) Watches seven corners of the web — YC batches, fresh funding, Hacker News "Who's Hiring", conference speakers, active GitHub contributors, employee updates, hiring tweets — and adds new companies the moment they start hiring. Off by default. Flip it on in settings.
Principles¶
-
Your laptop, your data
One user, one machine. Postgres and Redis run in Docker right next to the app. No shared server, no telemetry, no analytics.
-
You bring the API key
Nobody proxies your requests. Hook up your own OpenRouter key, or a local model via Ollama, or anything else you can wire in.
-
Opinionated
Your profile drives the scoring. Skills you do, skills you don't, salary floor, visa status, dealbreakers — the AI respects all of it.
-
Tweak it live
Blacklists, sources, thresholds, AI instructions, profile, model choice — all editable from the dashboard. Changes kick in on the next batch.
A note on scope¶
HireEx is not:
- a paid service — it's a tool you run for yourself.
- a recruiter tool — it's the candidate's side.
- a keyword matcher — the AI reads your positioning and the posting together, and respects your dealbreakers.
- a firehose — it aggressively archives near-misses and duplicates so your inbox stays small.
It is the tool I built because I got tired of scoring 400 postings a week by hand.