How siddanth-6365/vercellite works
Project type and purpose
• This is a full-stack “serverless” deployment platform prototype (“VercelLite”) with four main services:
• A Next.js frontend (UI)
• A Node.js/Express backend API server
• A build worker service (“build-servers”) that runs Docker builds
• A reverse-proxy service for custom domain routing
• Its goal is to let users submit a Git URL, build inside Docker/ECS, store artifacts in S3, and serve them via custom domains, with real-time build logs over WebSockets.
File structure
• Top-level directories:
• ui/ → Next.js frontend (app/, components/, hooks/, etc.)
• node-api-server/ → Express API, Prisma for Postgres, ECS/Redis/SOCKET integration
• build-servers/ → Shell scripts, Node.js scripts to clone, build, push to ECR/S3
• reverse-proxy-s3/ → Small Node.js reverse-proxy that maps custom domains to S3 buckets
• Configuration and infra:
• Dockerfile in build-servers/
• main.sh and script.js orchestrate the build container
• Prisma schema and migrations under node-api-server/prisma
• Separately defined package.json in each service
• Patterns:
• Microservices (each directory is an independent deployable)
• Event-driven build logs via Redis Pub/Sub
• Real-time client updates via Socket.IO
• Separation of concerns: UI ↔ API ↔ Builder ↔ Proxy
Extract key points from README
• Workflow: submit URL → build in Docker/ECS → store artifacts in S3 → map custom domain → stream logs via Redis + Socket.IO
• Tech stack calls out AWS S3, ECS/ECR, Redis, Socket.IO
• Prototype status: no advanced scaling/auth yet
Final advice
• Data flows end-to-end: from user action to artifact delivery
• Both control flows (API triggers) and data flows (logs, artifacts)
• Avoid clutter: group related sub-components (e.g., Prisma + Postgres) inside a single “Database” boundary
• Review the file tree to ensure no component is missed (e.g., reverse-proxy-s3, build-servers scripts)