┌─┐┌┬┐┌─┐┌─┐┬┌─┌─┐┬  ┌─┐┬ ┬
  └─┐ │ ├─┤│  ├┴┐├┤ │  │ ││││
  └─┘ ┴ ┴ ┴└─┘┴ ┴└  ┴─┘└─┘└┴┘

Build, run, and monitor AI-powered workflows — visually. StackFlow gives you a drag-and-drop graph editor, a LangGraph execution engine, and full Langfuse observability, all wired together with a one-command setup.

---

Get started

macOS

curl -fsSL https://raw.githubusercontent.com/your-org/StackFlow/main/install.sh | bash
stackflow start

Windows

Set-ExecutionPolicy -Scope CurrentUser -ExecutionPolicy RemoteSigned   # one-time
.\install.ps1
stackflow start

Open http://localhost:5173 and start building.

The install script handles everything: Python, Node.js, Docker, PostgreSQL, Langfuse — no manual setup required.

---

What's inside

• 🎨 Visual editor — Drag-and-drop graph editor with subgraphs, live execution status, undo/redo, and a variables panel
• 🤖 LangGraph engine — Stateful execution with PostgreSQL checkpointing and automatic resume on failure
• 🧩 Module system — Install integrations at runtime; each ships its own nodes, dependencies, and credentials
• 📡 Langfuse observability — Every execution is fully traced with spans, inputs, and outputs

---

Modules

Install the integrations you need:

stackflow pm list                 # see what's available
stackflow pm install slack        # install a module
stackflow pm install all          # install everything

Module	Nodes
AWS	EC2 security groups
Git	Branch management, commits
GitHub	Actions, file fetching, workflow monitoring
Jira	Ticket creation and management
Kubernetes	Pod logs, deployment monitoring
LLM	Repo-aware agents (OpenAI, Ollama, Gemini, and more)
Pull Requests	Full PR lifecycle — create, review, resolve, categorize
Slack	Messages, reactions, routing, assistant flows

---

CLI

stackflow start              # start everything (returns immediately)
stackflow stop               # stop API + editor
stackflow stop all           # stop everything including Docker infra
stackflow restart            # restart API server
stackflow restart ui         # restart editor
stackflow logs               # last 50 lines of API logs
stackflow logs ui -f         # follow UI logs
stackflow docker             # Docker container status
stackflow pm <command>       # package manager

---

Services

Service	URL
Editor	http://localhost:5173
API	http://localhost:8000
Langfuse	http://localhost:3000  ·  admin@stackflow.local / adminadmin

---

Requirements

• macOS — Homebrew. Everything else (Python 3.11+, Node.js, Docker via Colima) is installed by ./install.sh.
• Windows — Windows 10/11 with winget. Everything else (Python 3.11+, Node.js, Docker Desktop) is installed by .\install.ps1.