Dev Servers

SnakeFlow can manage any number of dev servers — Next.js, Vite, FastAPI, Django, Rails, Go, Rust, PHP — any runtime, any command.

Each server runs in a named VS Code terminal and is monitored via TCP port check + terminal presence. Status is visible in the sidebar and status bar in real time.

Server Configuration

Add servers to .vscode/settings.json:

"devManager.project.servers": [
  {
    "id": "web",
    "label": "Frontend",
    "command": "npm run dev",
    "path": "",
    "port": 3000
  },
  {
    "id": "api",
    "label": "API Server",
    "command": "python manage.py runserver 8000",
    "path": "backend",
    "port": 8000
  },
  {
    "id": "studio",
    "label": "Prisma Studio",
    "command": "npx prisma studio",
    "path": "",
    "port": 5555,
    "startWithAll": false
  }
]

Server Fields

Field	Required	Description
id	✅	Unique identifier — used as terminal name prefix
label	✅	Display name in the server menu and status bar
command	✅	Shell command to start the server
path	—	Relative working directory from project root ("" = root)
port	—	TCP port to monitor for live status (0 = no monitoring)
startWithAll	—	Include in Start all (Ctrl+Alt+S). Default true. Set false for optional tools you start only from the server menu (e.g. Prisma Studio, Storybook).

Commands

Action	Keybinding	How to open
Start all (bulk)	Ctrl+Alt+S	Tools menu (Ctrl+Alt+M) → Start all servers — only entries with startWithAll not false
Stop all	Ctrl+Alt+Q	Tools menu or server menu → Stop all — stops every configured server
Restart all	Ctrl+Alt+R	Server menu → Restart all — stop all, then bulk start
Per-server Start / Stop / Restart	—	Click the server indicator in the status bar → pick a server → choose action
Manage Servers	—	Command Palette → SnakeFlow: Manage Servers, or server menu → Manage Servers

Same idea as Containers

Container actions live in devManager.project.containerActions and run once from the Containers menu. Dev servers are long-running processes with port/terminal tracking — configure them in devManager.project.servers and use the status bar menu for one-off tools without adding hardcoded commands to the extension.

Examples

Single-app (Node.js)

[{ "id": "web", "label": "Dev Server", "command": "npm run dev", "path": "", "port": 3000 }]

Monorepo (Turborepo / pnpm)

[
  { "id": "web", "label": "Frontend", "command": "pnpm run dev", "path": "apps/web", "port": 5173 },
  { "id": "api", "label": "API", "command": "pnpm run dev", "path": "apps/api", "port": 8787 },
  {
    "id": "inngest",
    "label": "Inngest",
    "command": "pnpm dlx inngest-cli@latest dev",
    "path": "apps/web",
    "port": 8288
  }
]

Python + React

[
  { "id": "frontend", "label": "React", "command": "npm run dev", "path": "frontend", "port": 3000 },
  { "id": "backend", "label": "Django", "command": "python manage.py runserver 8000", "path": "backend", "port": 8000 },
  {
    "id": "celery",
    "label": "Celery Worker",
    "command": "celery -A config worker -l info",
    "path": "backend",
    "port": 0
  }
]

Go microservices

[
  {
    "id": "gateway",
    "label": "API Gateway",
    "command": "go run ./cmd/gateway",
    "path": "services/gateway",
    "port": 8080
  },
  { "id": "auth", "label": "Auth", "command": "go run ./cmd/auth", "path": "services/auth", "port": 8081 }
]

Root proxy pattern

If your root package.json has a proxy script like "dev": "npm run dev --prefix application", set "path": "" (root), not "application". The proxy runs from the root — pointing to the subdirectory would bypass it.

Optional tools (menu only)

Add a separate server entry for each command you want in the menu — nothing is built into the extension:

{
  "id": "studio",
  "label": "Prisma Studio",
  "command": "npx prisma studio",
  "path": "",
  "port": 5555,
  "startWithAll": false
}

• Ctrl+Alt+S skips entries with startWithAll: false
• Status bar → server menu → pick Prisma Studio → Start
• Stop all still stops Studio if it was started (same full stop cycle: Ctrl+C → port kill → dispose terminal)

When adding a server via Manage Servers, step 5 asks whether to include it in bulk start.

Status Monitoring

Server status is shown in:

• Status bar — combined server indicator; click to open the per-server menu
• Server menu — live status line per entry (Running, Stopped, port/PID when applicable)

Status uses a combined check: TCP port open AND the terminal is still running. This avoids false positives from external processes using the same port.

Status	Meaning
🟢 Running	Port is open and terminal is alive
🔴 Stopped	Port is closed or terminal was closed
⚪ No port	Server has port: 0 — terminal presence only