Show HN: OpenScreen. Open-source video assessment screening tool (github.com)

OpenScreen is an open-source video screening platform released on Show HN that automates scoring and feedback for video responses using optional AI. Aimed at recruitment, education, and training workflows, it offers customizable campaigns and rubrics plus AI-generated scoring via Google Gemini 2.5, letting teams mix human-reviewed rubrics with automated assessment. The product includes a modern Next.js/React UI, a pay-per-use credit model (Stripe integration), and built-in auth and privacy controls—positioning it as a flexible, self-hostable alternative to closed commercial video-assessment tools. Technically, OpenScreen is a Next.js 15 + React 19 frontend with Tailwind, Radix UI components, and a Firebase backend (Auth, Firestore, Storage, Cloud Functions) for serverless scaling; Cloud Functions include Firestore triggers for AI processing. Requirements include Node.js 20+, a Firebase project on the Blaze plan for production, a Google Gemini API key, and optionally a Stripe account. The repo is AGPL-licensed and structured for easy local dev and deployment to Firebase Hosting. Important trade-offs: quick setup and extensibility for teams that want control, but reliance on Gemini, Firebase billing, and third-party payments means ongoing API costs and vendor dependency to consider.