diff --git a/framework/cmd.d/node b/framework/cmd.d/node
index 05f3b98..c412659 100755
--- a/framework/cmd.d/node
+++ b/framework/cmd.d/node
@@ -4,4 +4,4 @@ set -eu
 
 DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 
-"$DIR"/../shims/node "$@"
+exec "$DIR"/../shims/node "$@"
diff --git a/master/main.go b/master/main.go
index 2b011da..2aaa94c 100644
--- a/master/main.go
+++ b/master/main.go
@@ -1,23 +1,33 @@
 package main
 
 import (
-	// "context"
 	"fmt"
 	"os"
 	"os/signal"
-	// "sync"
+	"strconv"
 	"syscall"
 )
 
 func main() {
-	//	 var	program1 = os.Getenv("BUILD_COMMAND")
-	//var	program2 = os.Getenv("RUN_COMMAND")
+	watchedDir := os.Getenv("WATCHED_DIR")
 
-	var watchedDir = os.Getenv("WATCHED_DIR")
+	numChildProcesses := 1
+	if n, err := strconv.Atoi(os.Getenv("NUM_CHILD_PROCESSES")); err == nil && n > 0 {
+		numChildProcesses = n
+	}
 
-	// Create context for graceful shutdown
-	//	ctx, cancel := context.WithCancel(context.Background())
-	//defer cancel()
+	basePort := 3000
+	if p, err := strconv.Atoi(os.Getenv("BASE_PORT")); err == nil && p > 0 {
+		basePort = p
+	}
+
+	listenPort := 8080
+	if p, err := strconv.Atoi(os.Getenv("LISTEN_PORT")); err == nil && p > 0 {
+		listenPort = p
+	}
+
+	// Create worker pool
+	pool := NewWorkerPool()
 
 	// Setup signal handling
 	sigCh := make(chan os.Signal, 1)
@@ -27,24 +37,16 @@ func main() {
 
 	go watchFiles(watchedDir, fileChanges)
 
-	go runExpress(fileChanges)
+	go runExpress(fileChanges, numChildProcesses, basePort, pool)
 
-	// WaitGroup to track both processes
-	// var wg sync.WaitGroup
-
-	// Start both processes
-	//wg.Add(2)
-	//	 go runProcess(ctx, &wg, "builder", program1)
-	// go runProcess(ctx, &wg, "runner", program2)
+	// Start the reverse proxy
+	listenAddr := fmt.Sprintf(":%d", listenPort)
+	go startProxy(listenAddr, pool)
 
 	// Wait for interrupt signal
 	<-sigCh
 	fmt.Println("\nReceived interrupt signal, shutting down...")
 
-	// Cancel context to signal goroutines to stop
-	///	cancel()
 
-	// Wait for both processes to finish
-	// wg.Wait()
 	fmt.Println("All processes terminated cleanly")
 }
diff --git a/master/proxy.go b/master/proxy.go
new file mode 100644
index 0000000..1ee52b2
--- /dev/null
+++ b/master/proxy.go
@@ -0,0 +1,48 @@
+package main
+
+import (
+	"log"
+	"net/http"
+	"net/http/httputil"
+	"net/url"
+)
+
+// startProxy starts an HTTP reverse proxy that forwards requests to workers.
+// It acquires a worker from the pool for each request and releases it when done.
+func startProxy(listenAddr string, pool *WorkerPool) {
+	handler := http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		// Acquire a worker (blocks if none available)
+		workerAddr, ok := pool.Acquire()
+		if !ok {
+			http.Error(w, "Service unavailable", http.StatusServiceUnavailable)
+			return
+		}
+
+		// Ensure we release the worker when done
+		defer pool.Release(workerAddr)
+
+		// Create reverse proxy to the worker
+		targetURL, err := url.Parse("http://" + workerAddr)
+		if err != nil {
+			log.Printf("[proxy] Failed to parse worker URL %s: %v", workerAddr, err)
+			http.Error(w, "Internal server error", http.StatusInternalServerError)
+			return
+		}
+
+		proxy := httputil.NewSingleHostReverseProxy(targetURL)
+
+		// Custom error handler
+		proxy.ErrorHandler = func(w http.ResponseWriter, r *http.Request, err error) {
+			log.Printf("[proxy] Error proxying to %s: %v", workerAddr, err)
+			http.Error(w, "Bad gateway", http.StatusBadGateway)
+		}
+
+		log.Printf("[proxy] %s %s -> %s", r.Method, r.URL.Path, workerAddr)
+		proxy.ServeHTTP(w, r)
+	})
+
+	log.Printf("[proxy] Listening on %s", listenAddr)
+	if err := http.ListenAndServe(listenAddr, handler); err != nil {
+		log.Fatalf("[proxy] Failed to start: %v", err)
+	}
+}
diff --git a/master/runexpress.go b/master/runexpress.go
index f5fac71..a2333f8 100644
--- a/master/runexpress.go
+++ b/master/runexpress.go
@@ -1,6 +1,7 @@
 package main
 
 import (
+	"fmt"
 	"io"
 	"log"
 	"os"
@@ -10,43 +11,45 @@ import (
 	"time"
 )
 
-func runExpress(changes <-chan FileChange) {
-	var currentProcess *exec.Cmd
+func runExpress(changes <-chan FileChange, numProcesses int, basePort int, pool *WorkerPool) {
+	var currentProcesses []*exec.Cmd
 	var mu sync.Mutex
 
-	// Helper to start the express process
-	startExpress := func() *exec.Cmd {
-		cmd := exec.Command("../express/run.sh")
+	// Helper to start an express process on a specific port
+	startExpress := func(port int) *exec.Cmd {
+		listenAddr := fmt.Sprintf("127.0.0.1:%d", port)
+		cmd := exec.Command("../express/run.sh", "--listen", listenAddr)
 		cmd.Stdout = os.Stdout
 		cmd.Stderr = os.Stderr
 
 		if err := cmd.Start(); err != nil {
-			log.Printf("[express] Failed to start: %v", err)
+			log.Printf("[express:%d] Failed to start: %v", port, err)
 			return nil
 		}
 
-		log.Printf("[express] Started (pid %d)", cmd.Process.Pid)
+		log.Printf("[express:%d] Started (pid %d)", port, cmd.Process.Pid)
 
 		// Monitor the process in background
-		go func() {
+		go func(p int) {
 			err := cmd.Wait()
 			if err != nil {
-				log.Printf("[express] Process exited: %v", err)
+				log.Printf("[express:%d] Process exited: %v", p, err)
 			} else {
-				log.Printf("[express] Process exited normally")
+				log.Printf("[express:%d] Process exited normally", p)
 			}
-		}()
+		}(port)
 
 		return cmd
 	}
 
-	// Helper to stop the express process
+	// Helper to stop an express process
 	stopExpress := func(cmd *exec.Cmd) {
 		if cmd == nil || cmd.Process == nil {
 			return
 		}
 
-		log.Printf("[express] Stopping (pid %d)", cmd.Process.Pid)
+		pid := cmd.Process.Pid
+		log.Printf("[express] Stopping (pid %d)", pid)
 		cmd.Process.Signal(syscall.SIGTERM)
 
 		// Wait briefly for graceful shutdown
@@ -58,13 +61,38 @@ func runExpress(changes <-chan FileChange) {
 
 		select {
 		case <-done:
-			log.Printf("[express] Stopped gracefully")
+			log.Printf("[express] Stopped gracefully (pid %d)", pid)
 		case <-time.After(5 * time.Second):
-			log.Printf("[express] Force killing")
+			log.Printf("[express] Force killing (pid %d)", pid)
 			cmd.Process.Kill()
 		}
 	}
 
+	// Helper to stop all express processes
+	stopAllExpress := func(processes []*exec.Cmd) {
+		for _, cmd := range processes {
+			stopExpress(cmd)
+		}
+	}
+
+	// Helper to start all express processes and update the worker pool
+	startAllExpress := func() []*exec.Cmd {
+		processes := make([]*exec.Cmd, 0, numProcesses)
+		addresses := make([]string, 0, numProcesses)
+		for i := 0; i < numProcesses; i++ {
+			port := basePort + i
+			addr := fmt.Sprintf("127.0.0.1:%d", port)
+			cmd := startExpress(port)
+			if cmd != nil {
+				processes = append(processes, cmd)
+				addresses = append(addresses, addr)
+			}
+		}
+		// Update the worker pool with new worker addresses
+		pool.SetWorkers(addresses)
+		return processes
+	}
+
 	// Helper to run the build
 	runBuild := func() bool {
 		log.Printf("[build] Starting ncc build...")
@@ -97,6 +125,14 @@ func runExpress(changes <-chan FileChange) {
 	var debounceTimer *time.Timer
 	const debounceDelay = 100 * time.Millisecond
 
+	// Initial build and start
+	log.Printf("[master] Initial build...")
+	if runBuild() {
+		currentProcesses = startAllExpress()
+	} else {
+		log.Printf("[master] Initial build failed")
+	}
+
 	for change := range changes {
 		log.Printf("[watch] %s: %s", change.Operation, change.Path)
 
@@ -107,18 +143,18 @@ func runExpress(changes <-chan FileChange) {
 
 		debounceTimer = time.AfterFunc(debounceDelay, func() {
 			if !runBuild() {
-				log.Printf("[master] Build failed, keeping current process")
+				log.Printf("[master] Build failed, keeping current processes")
 				return
 			}
 
 			mu.Lock()
 			defer mu.Unlock()
 
-			// Stop old process
-			stopExpress(currentProcess)
+			// Stop all old processes
+			stopAllExpress(currentProcesses)
 
-			// Start new process
-			currentProcess = startExpress()
+			// Start all new processes
+			currentProcesses = startAllExpress()
 		})
 	}
 }
diff --git a/master/watchfiles.go b/master/watchfiles.go
index b339def..c7d31b3 100644
--- a/master/watchfiles.go
+++ b/master/watchfiles.go
@@ -1,12 +1,32 @@
 package main
 
 import (
-	"github.com/fsnotify/fsnotify"
 	"log"
 	"os"
 	"path/filepath"
+	"strings"
+
+	"github.com/fsnotify/fsnotify"
 )
 
+// shouldIgnore returns true for paths that should not trigger rebuilds
+func shouldIgnore(path string) bool {
+	// Ignore build output and dependencies
+	ignoreDirs := []string{"/dist/", "/node_modules/", "/.git/"}
+	for _, dir := range ignoreDirs {
+		if strings.Contains(path, dir) {
+			return true
+		}
+	}
+	// Also ignore if path ends with these directories
+	for _, dir := range []string{"/dist", "/node_modules", "/.git"} {
+		if strings.HasSuffix(path, dir) {
+			return true
+		}
+	}
+	return false
+}
+
 func watchFiles(dir string, changes chan<- FileChange) {
 	watcher, err := fsnotify.NewWatcher()
 	if err != nil {
@@ -14,12 +34,15 @@ func watchFiles(dir string, changes chan<- FileChange) {
 	}
 	defer watcher.Close()
 
-	// Add all directories recursively
+	// Add all directories recursively (except ignored ones)
 	err = filepath.Walk(dir, func(path string, info os.FileInfo, err error) error {
 		if err != nil {
 			return err
 		}
 		if info.IsDir() {
+			if shouldIgnore(path) {
+				return filepath.SkipDir
+			}
 			err = watcher.Add(path)
 			if err != nil {
 				log.Printf("Error watching %s: %v\n", path, err)
@@ -38,6 +61,11 @@ func watchFiles(dir string, changes chan<- FileChange) {
 				return
 			}
 
+			// Skip ignored paths
+			if shouldIgnore(event.Name) {
+				continue
+			}
+
 			// Handle different types of events
 			var operation string
 			switch {
diff --git a/master/workerpool.go b/master/workerpool.go
new file mode 100644
index 0000000..e35a6f7
--- /dev/null
+++ b/master/workerpool.go
@@ -0,0 +1,75 @@
+package main
+
+import (
+	"log"
+	"sync"
+)
+
+// WorkerPool manages a pool of worker addresses and tracks their availability.
+// Each worker can only handle one request at a time.
+type WorkerPool struct {
+	mu        sync.Mutex
+	workers   []string
+	available chan string
+	closed    bool
+}
+
+// NewWorkerPool creates a new empty worker pool.
+func NewWorkerPool() *WorkerPool {
+	return &WorkerPool{
+		available: make(chan string, 100), // buffered to avoid blocking
+	}
+}
+
+// SetWorkers updates the pool with a new set of worker addresses.
+// Called when workers are started or restarted after a rebuild.
+func (p *WorkerPool) SetWorkers(addrs []string) {
+	p.mu.Lock()
+	defer p.mu.Unlock()
+
+	// Drain the old available channel
+	close(p.available)
+	for range p.available {
+		// drain
+	}
+
+	// Create new channel and populate with new workers
+	p.available = make(chan string, len(addrs)+10)
+	p.workers = make([]string, len(addrs))
+	copy(p.workers, addrs)
+
+	for _, addr := range addrs {
+		p.available <- addr
+	}
+
+	log.Printf("[pool] Updated workers: %v", addrs)
+}
+
+// Acquire blocks until a worker is available and returns its address.
+func (p *WorkerPool) Acquire() (string, bool) {
+	addr, ok := <-p.available
+	if ok {
+		log.Printf("[pool] Acquired worker %s", addr)
+	}
+	return addr, ok
+}
+
+// Release marks a worker as available again after it finishes handling a request.
+func (p *WorkerPool) Release(addr string) {
+	p.mu.Lock()
+	defer p.mu.Unlock()
+
+	// Only release if the worker is still in our current set
+	for _, w := range p.workers {
+		if w == addr {
+			select {
+			case p.available <- addr:
+				log.Printf("[pool] Released worker %s", addr)
+			default:
+				// Channel full, worker may have been removed
+			}
+			return
+		}
+	}
+	// Worker not in current set (probably from before a rebuild), ignore
+}