Open Source • MIT License

MULTI
TASKING,
FINALLY.

Turn any task into parallelized AI workflows with 176+ curated skills. Say what you want built. WinDAGs figures out how.

View Source
100% FreeLocal-FirstZero Config
WinDAGs - Control Center
> Build me a landing page for CloudMetrics
WAVE EXECUTION
Wave 0
research-analyst
Wave 1
web-design-expert
competitive-cartographer
Wave 2
design-system-creator
technical-writer
Wave 3
fullstack-debugger

HOW IT WORKS

From natural language to parallel execution in four steps.

01

DESCRIBE

Tell WinDAGs what you want in plain English. No jargon required.

02

DECOMPOSE

Your request gets broken into subtasks automatically. DAGs form.

03

MATCH

Each subtask gets matched to specialist skills from our 176+ library.

04

EXECUTE

Agents run in parallel waves. Outputs flow forward. You stay in control.

BUILT FOR POWER USERS

The intelligence layer for Claude Code. No more one-task-at-a-time.

176+ Curated Skills

From API architecture to crisis response. Each skill makes an agent an instant expert. Quality over quantity.

dag-executor, api-architect, grief-companion, recovery-coach...
Wave-Based Parallelism

Nodes with resolved dependencies run simultaneously. 5 waves of agent calls that feed forward into each other.

Wave 0 → Wave 1 → Wave 2 → Wave 3 → Done
Zero Token Overhead

Spawns `claude -p` CLI processes—not nested API calls. Each agent runs independently with skill-injected expertise.

ProcessExecutor • WorktreeExecutor • TaskToolExecutor
Human-in-the-Loop

Approval gates let you review before execution. Cost estimates before you spend. Stop any time.

Review → Approve → Execute → Celebrate

176+ SKILLS
ONE TEAM

Skills are expertise packages. They get injected into agents at runtime, turning generic Claude into a specialist. You don't pick skills—WinDAGs matches them automatically.

Browse All Skills
24
DAG Orchestration
18
Code Quality
22
Design & Creative
12
Recovery & Health
15
DevOps & SRE
14
AI & ML
8
Legal & Compliance
63
And more...

WHY WINDOWS 3.11?

We didn't pick Windows 3.1 randomly. Windows for Workgroups was Microsoft's first attempt at cooperative multitasking and network-aware computing.

WinDAGs is the same idea, 30 years later, for AI agents.

Workgroups = The DAGs themselves (skills collaborating)
Win = DAGs for the win
Cooperative multitasking = Wave-based parallel execution

It's not a joke. It's a statement.

STAY IN THE LOOP

Get notified when we launch new features, skills, and integrations. No spam. Unsubscribe anytime.

📧 Email Signup