HRCAN I VIBECODE
Greenhouse?
BUILD FIRST
- A non-sensitive or read-only version of the core experience
- A manual review step wherever correctness or trust matters
- A narrow companion tool that removes a real piece of work
LEAVE FOR LATER
- Regulated decisions, payments, or irreversible actions
- Claims of correctness, compliance, or uptime that the MVP cannot support
- Sensitive integrations until security, support, and auditability are ready
RESEARCH NOTES
What this tool actually does.
Greenhouse is a commercial applicant tracking system (ATS) and hiring platform. It structures recruiting around configurable job pipelines, interview kits (stages, questions, scorecard attributes), and collaborative scorecard-based evaluation. It includes resume parsing, duplicate/spam detection, talent rediscovery, branded job boards, and a candidate portal (MyGreenhouse) with real-time status updates. AI features (OpenAI-backed, tier-dependent) automate hiring plan generation, interview question suggestions, scorecard summaries, report filter suggestions, job note summarization, prospect response review, and offer forecasting; the Pro tier adds resume anonymization to reduce bias. Administration includes automated tasks, role-based permissions, governance, analytics, and hundreds of pre-built integrations. Data is encrypted in transit and at rest.
RESEARCHED BUILD PROMPT
Give this to your coding agent.
Build a web-based applicant tracking system (a "Greenhouse-like structured hiring platform"). Core workflow: hiring managers create a job → recruiters design a hiring plan from reusable interview kits (stages, per-stage questions, weighted scorecard attributes) → publish a branded job board → candidates apply via a public portal → resumes are parsed and displayed with optional PII anonymization → recruiters review applications, filter by keyword/attribute, and move candidates through stages → interviewers submit scorecards → the system summarizes agreements/disagreements per candidate → recruiters send offers → dashboards highlight pipeline bottlenecks with natural-language report filters.
User-facing requirements: recruiter/admin dashboard (jobs, candidates, kanban pipeline), interview-kit builder, scorecard forms, candidate portal showing application status, reporting UI with prompt-to-filter, role-based permissions (admin/recruiter/interviewer/hiring manager), and an audit log.
Technical requirements: PostgreSQL schema for jobs, candidates, applications, stages, interviews, scorecards, and offers; REST API; background queue for resume parsing and AI tasks; LLM integration for interview-question suggestions, scorecard summaries, and report-filter parsing; encryption at rest and in transit; seeded demo data.
Implementation approach: Next.js + TypeScript frontend, FastAPI backend, PostgreSQL, Celery/Redis for async workers, OpenAI-compatible LLM client, and a resume-parsing library (e.g., Apache Tika). Use a responsive plain UI with an original design system — do not copy Greenhouse's brand, logo, colors, or exact layout.
Out of scope: identity verification, candidate fraud detection, native mobile apps, enterprise SSO, a third-party integration marketplace, and multi-tenant billing.
Deliver a README covering setup, architecture overview, and important limitations (e.g., LLM output nondeterminism, parser accuracy, single-tenant by default).