What it does

JustHireMe is a Tauri-based desktop app with a Python FastAPI sidecar that turns job hunting into a local data pipeline. It scrapes listings from job boards, RSS feeds, and community sources; runs them through a deterministic quality gate to drop stale or spammy posts; scores fit against your profile using a local ONNX embedding model; and generates tailored PDF resumes, cover letters, and outreach drafts. Everything lives on your machine—SQLite for CRM data, Kuzu for profile graphs, LanceDB for vectors.

The interesting bit

The “thin installer” trick is genuinely clever: the download is only ~100 MB because the heavy stuff (Playwright browser, vector libraries, embedding model) gets pulled once on first run and cached. The app also auto-updates from GitHub releases. It is a pragmatic solution to the “local-first but still usable” problem that usually forces a 700 MB upfront download.

Key highlights

• AGPL-licensed, ships for Windows, macOS, and Linux with auto-updates
• Local semantic matching via all-MiniLM-L6-v2—no API key required for core ranking
• Explainable scoring: deterministic rules plus optional LLM reasoning, not a black box
• Profile-aware matching using Kuzu graph data and LanceDB vectors, not keyword brute force
• Browser automation and auto-apply exist but are explicitly experimental and disabled by default

Caveats

• macOS builds are ad-hoc signed but not notarized, so Gatekeeper will likely complain
• The heavy runtime downloads on first run—plan for that if you are on a slow or metered connection
• OS keychain integration for API keys is listed as “planned,” not shipped

Verdict

Worth a look if you are privacy-sensitive about your job search or tired of SaaS tools that spray your resume everywhere. Skip it if you want a fully managed cloud service or need mobile access; this is a desktop workbench, not an app you check on the train.