kestrel-app 0.12.0

AI-powered job search platform. Self-hosted, open-source, privacy-first.

Kestrel

Stop spray-and-praying. Start running your job search like a system.
Kestrel scrapes job boards, scores every posting against your profile, and tracks your pipeline — all on your machine.

---

Pipeline — drag applications across stages

Discovery — AI-scored job matches

---

What it does

You tell Kestrel what you're looking for. It does the rest.

• Scrapes job boards — Indeed, LinkedIn, Glassdoor, Arbeitsagentur. Runs daily. You wake up to new matches.
• Scores every posting — AI reads the job, compares it to your profile, and gives you a fit score (0-10) and a desire score (0-10). Two numbers, because "can you do it?" and "do you want it?" are different questions.
• Tracks your pipeline — Kanban board. Drag jobs between stages. See where everything stands.
• Preps you for interviews — Company research, mock questions, a STAR story library you build over time.
• Runs on your machine — Your data stays on your disk. No account, no cloud, no tracking. Unless you choose otherwise.

Demo Mode works offline, for free, right now. Real AI scoring costs $1-5/month with the built-in optimizations (prompt caching, model routing, response caching — 8 techniques stacked).

Why Kestrel

Job search tools fall into two camps. The SaaS apps (Huntr, Teal, Simplify) want $30/month and your data on their servers. The open-source tools are either abandoned or developer-only CLI scripts.

Kestrel is neither. It's a full app — web dashboard, REST API, mobile app (in progress) — that runs locally. You own the data, pick your own AI provider (or use none), and pay nothing for the software itself.

Full comparison →

Install

Pick one. They all give you the same app.

# One command (detects OS, installs, opens browser)
curl -fsSL https://raw.githubusercontent.com/pleasedodisturb/kestrel/main/install.sh | bash

# Or with pip
pip install kestrel-app && kestrel start

# Or with Node.js
npx kestrel-app

# Or with Homebrew
brew install pleasedodisturb/kestrel/kestrel && kestrel start

# Or with Docker (nothing touches your system)
git clone https://github.com/pleasedodisturb/kestrel.git && cd kestrel && bash setup.sh

Zero install option: Open in GitHub Codespaces — your own instance in 2 minutes, free with a GitHub account.

Need hand-holding? The step-by-step guide assumes you've never used a terminal.

Add AI (optional)

Kestrel works without AI. Demo Mode gives you the full UI with simulated scores — enough to explore everything and decide if it's for you.

When you're ready for real scoring: go to Settings, click "Connect to OpenRouter," log in. Done. Free-tier models (Llama 3.3 70B) score jobs at zero cost.

Want more control? Kestrel supports 6 providers — from free local models to the best commercial APIs. You can even run multiple at once.

Provider	Cost	Privacy	Good for
Demo Mode	Free	Perfect	Trying it out
OpenRouter	$0-10/mo	Good	Most people — one account, 300+ models
Anthropic	$1-10/mo	Excellent (7-day retention)	Best quality, prompt caching
Together AI	$1-5/mo	Good (ZDR available)	Bulk scoring on a budget
Ollama	Free	Perfect	Nothing leaves your machine
OpenAI	$1-10/mo	Good	GPT models direct

Full guide with the electricity analogy → · Technical comparison + API keys → · Privacy deep dive →

Docs

Quickstart	First-time setup, zero assumptions
FAQ	Common questions answered
Help	Something broke? Start here
How Scoring Works	What "fit score" means, how Kestrel decides
How Testing Works	2,800+ automated checks explained
Comparison	Kestrel vs Huntr, Teal, Simplify, and others
All docs →	Full index of 40+ documents

Contributing

Contributing guide · Deployment · API Reference

License

AGPL-3.0 — free and open source. If you modify Kestrel and offer it as a service, share your changes under the same license.