Job-search automation you run yourself, and a public feed of the job listings it finds.
Job Radar is a self-hosted tool that scrapes career pages across 100+ companies (Workday, Greenhouse, SmartRecruiters, SAP SuccessFactors, and more), scores matches against your own profile, and drafts cover letters and tailored CVs with an LLM -- with a human reviewing everything before anything is ever sent.
The whole personal pipeline: scrapers, matching, letter/CV generation, a review dashboard. Designed to run comfortably on something as small as a Raspberry Pi. MIT-licensed. Nothing is ever sent automatically -- you approve every step.
A separate, tokengated read API serving the generic listing data a Job Radar instance finds -- title, company, location, URL, description. No personal data: no relevance scores, no application status, no letters or CVs. Fed by a scrape cycle that already knows how to talk to dozens of different ATS platforms, so you don't have to reverse-engineer them yourself.
This site is built to be readable by AI agents and other automated tools, not just browsers:
| Resource | Purpose |
|---|---|
| /llms.txt | Plain-Markdown site index |
| /openapi.json | OpenAPI 3.0 spec for the Job Radar Hub API |
| /.well-known/api-catalog | Machine-readable resource catalog (RFC 9727) |
| /.well-known/mcp/server-card.json | MCP server card |
| /robots.txt | Explicitly allows major AI crawlers |
The same data as the REST API, over the Model Context Protocol, for MCP-aware
agent clients. One tool: search_jobs (company, source, title/location
substring, active status -- same filters as GET /jobs).
Endpoint: http://136.65.194.129:5301/mcp
Transport: Streamable HTTP
Auth: Authorization: Bearer <your_api_key> (same keys as the REST API)Server card: /.well-known/mcp/server-card.json