Stats
Links
Categories
Engy
AI-assisted engineering workspace for spec-driven development.
What is Engy
Engy is a personal dev workspace for the Specify → Plan → Execute → Complete loop. Write specs, plan milestones, run AI agents against tasks, review diffs, and send inline feedback to Claude Code — all without leaving the app.
Everything is accessible to AI agents via a built-in MCP server, so Claude Code CLI running in your terminal can read and write Engy data directly.
Features
Project Planning & Management
Plan your projects and break them into milestones and task groups. Start implementing milestone, task groups or individual tasks with a single click (🔨 icon) while you continue working on other things.
Spec Editor and Review
Rich text editor for writing and reviewing project plans. Supports headings, tables, lists, code blocks, and @ file mentions. Leave comments directly on any markdown file and send straight to a running Claude Code terminal session, so your AI agent gets feedback without you leaving the editor.
Task Management
Three views for managing tasks:
- Kanban — Todo / In Progress / Review / Done columns
- Eisenhower Matrix — prioritize by urgency and importance
- Dependency Graph — visualize task dependencies across layers
Diff Review & Inline Comments
Review uncommitted changes and branch diffs with line-level commenting. Leave inline comments on specific lines and send feedback directly to a running Claude Code session.
\
Claude Plans
Review AI-generated plans and send structured feedback directly to a running Claude Code session.
Built in Terminal
Engy has a built-in terminal that you can use to interact with Claude Code and manage your project. It persists across sessons and project pages, so you can keep your Claude Code agent running while you navigate around the app.
DevContainers
Engy comes with a pre-configured DevContainer setup. Once enabled from Workspace settings it will create the necessary files in the .devcontainer/ directory in the workspace root. You can customize the container configuration and post-start script as needed. The default setup includes strict firewall rules to allow communication between the container and the host machine and some default websites.
MCP Server
Built-in Model Context Protocol server so AI agents (Claude Code CLI) can read specs, create tasks, and update project state directly.
Workspaces
Permanent homes for ongoing concerns, tied to one or more repositories where Claude Code runs implementations. Additional repos can be added as extra directories. Organize projects, tasks, docs, and memory under one roof.
Notifications
Get notified when a plan is ready for review or other events need your attention.
Skills
Engy ships as a Claude Code plugin with skills that drive the full development loop from your terminal.
Install the marketplace:
/plugin marketplace add <cloned repo path>
Install the plugin.
Then you can run skills like:
| Skill | What it does |
|---|---|
/write-spec | Create or validate an SRS from source documents |
/milestone-plan | Break a spec into milestones, task groups, tasks, and dependencies |
/plan | Write a plan for a single task |
/validate-plan | Check a plan against its parent spec for alignment and gaps |
/implement | Implement a single task |
/implement-milestone | Orchestrate an entire milestone across task groups in parallel |
/review | Code review with auto-detected scope and severity-tagged findings |
/workspace-assistant | Quick task tracking — log bugs, create one-off work items |
Getting Started
Work in progress — active development Expect rough edges, missing features, and things that break.
Prerequisites: Node.js 20+, pnpm 10+
Development
pnpm install
# Start both the web server and client daemon
pnpm dev
Production
pnpm install
pnpm build
# Optionally set up environment
cp .env.example .env
# Uses .env if available, otherwise defaults
pnpm start
Open http://localhost:3000 (or whichever port you configured).
The web/ server runs on the configured port. The client/ daemon connects to it over WebSocket — it handles local filesystem access and git operations on your machine.
Environment variables (see .env.example):