Cloud SaaS
Teal · Huntr · Jobscan
Powerful, but paid by the month and your whole search lives on someone else’s servers. How hiring works where you are is not its problem.
The augmentation principle
Drafting is automated.
Submitting is not.
An open-source, local-first desktop app for an AI-powered job search. Your data, your machine, your AI. It scores roles, tailors your CV, and tracks the pipeline, then hands every decision back to you.
The gap we fill
Three tools, one missing one.
CLI pipelines
career-ops
Full pipeline, local-first, free, and excellent. But it is terminal-only, so it speaks to developers and no one else.
Desktop, local, free
applye.dev
The full pipeline as a desktop GUI: local-first, free, MIT, and aware of how hiring actually works where you are. Setup in 3 minutes, not 15. No terminal required.
career-ops gives developers a CLI. Applye gives everyone a desktop.
What is Applye?
Applye is a desktop app that runs the whole job-search loop end to end, on your machine. You paste a job description; it gives you a blunt HR and ATS check; it drafts a tailored CV you review and export; you move the role across a pipeline kanban as it progresses; and it helps you prepare for the interview. It is local-first, you bring your own AI (a direct API key or the CLI subscription you already pay for) and it is MIT-licensed and free.
What it does
Built to give you signal, not encouragement.
Blunt recruiter check
Paste a job and get an honest fit score, the keywords you are missing, the red flags a screener would catch, and a plain ATS pass/fail, the way a recruiter actually reads in the first ten seconds.
→ No encouragement. Just signal.
How the scoring works →Tailored CV in three passes
An XYZ rewrite, then a dual critique that argues with itself, then a clean build exported as a PDF that survives ATS parsing. You read every line before it exists as a file.
→ Applye drafts. You review, export, and submit.
Pipeline as a kanban
Drag each role from saved to applied to interview to offer. Stages are auto-dated and overdue applications wear a badge, so nothing quietly goes cold.
→ Your board, on your machine, not a vendor dashboard.
Bring your own AI
Plug in a direct API key - Anthropic or DeepSeek - or bridge a CLI subscription you already pay for: Claude Code or Codex. Code does the cheap, deterministic work; the model is only asked to judge.
→ Token-economical by design. A real search costs cents.
Local-first & private
Everything lives in one SQLite file on your disk. No account to create, no cloud to sync to, no telemetry phoning home. Delete the file and it is gone.
→ No cloud, no account, no tracking. Ever.
Works with the AI you already pay for.
Applye ships no model of its own and resells no tokens. Point it at a provider key, or bridge a CLI subscription you already have - the calls go straight from your machine to them.
Direct API keys
- Anthropic Claude
- DeepSeek
CLI subscriptions, bridged
- Claude Code
- Codex CLI
Independent trademarks of their owners. No affiliation or endorsement implied.
Local rules, handled
Made for the search you are actually running.
A job search is local even when the job is remote. Applye works anywhere, and where a market has its own conventions and paperwork it handles them instead of pretending everyone applies the same way.
- Documents in your language: CVs, cover letters, and interview prep in any of six languages, matched to what the role expects.
- Local conventions respected: photo or no photo, date and layout norms, and the ATS quirks that differ by market.
- Visa and work-permit awareness, including the EU Blue Card, for anyone applying across a border.
- GDPR-aligned by architecture: your data never leaves your machine, which satisfies the strictest regime by default.
- Germany, in depth: generate the Agentur für Arbeit Eigenbemühungen report straight from your tracked applications.
Local-first
One SQLite file on your machine.
Privacy by design
Nothing is collected. No telemetry.
Free / MIT
Open source, free.
Bring your own AI
Your key or your CLI subscription.
Augment, not automate
AI drafts. You decide and submit.
Open source & honest
Your data never leaves your machine.
Applye is MIT-licensed and developed in the open. Read the code, read the data guarantee, run it yourself.
When to use Applye
- You want fewer, better-targeted applications.
- You care where your job-search data lives.
- You already pay for an AI subscription or have an API key.
- You are applying across borders, or in a market with its own paperwork.
What Applye is not
- Not an auto-apply bot. It never submits for you.
- Not a scraper of closed boards. Discover reads public APIs and feeds; the rest you paste.
- Not a cloud service: no account, no server, no sync.
- Not a way to fake experience. Honesty over inflation.
FAQ
Straight answers.
You paste a job description; code extracts the requirements and Applye asks your AI to read it the way a recruiter or ATS would: a fit score, the missing keywords, and the red flags. The same job is never scored twice: results are cached against a hash of the text, so you do not pay tokens to re-read it.