The Challenge
Build a system that lets engineering teams define, visualize, and execute AI-powered workflows — from automated triage to code review — with both an open-source core and a commercial cloud offering.
What I Built
SWEny Core (Open Source): A workflow orchestration engine that executes DAG-based workflows using Claude Code. Built-in skills for GitHub, Linear, Sentry, Datadog, Slack, and BetterStack. Ships as a CLI tool and GitHub Action.
SWEny Studio: A visual DAG editor and live execution monitor built with React, XYFlow, and ELK graph layout. Supports undo/redo and real-time execution tracking.
SWEny Cloud (SaaS): Managed platform with team collaboration, job scheduling, analytics, and Stripe billing. Microservices backend with Hono API, BullMQ workers, and Drizzle ORM on PostgreSQL.
Technical Highlights
- Full monorepo with clear package boundaries (core, studio, cloud API, cloud UI, cloud worker)
- Claude Agent SDK integration for AI-powered task execution
- Multi-tenant architecture with Clerk auth
- Real-time job monitoring via WebSocket