unaiverse 0.1.12__py3-none-any.whl

unaiverse/networking/p2p/lib.go

@@ -0,0 +1,2765 @@
+// lib.go
+// This Go program compiles into a C shared library (.so file on Linux/macOS, .dll on Windows)
+// exposing libp2p functionalities (host creation, peer connection, pubsub, direct messaging)
+// for use by other languages, primarily Python via CFFI or ctypes.
+package main
+
+/*
+#include <stdlib.h>
+*/
+import "C" // Enables CGo features, allowing Go to call C code and vice-versa.
+
+import (
+	// Standard Go libraries
+	"bytes"           // For byte buffer manipulations (e.g., encoding/decoding, separators)
+	"errors"          // For handling some types of errors
+	"container/list"  // For an efficient ordered list (doubly-linked list for queues)
+	"context"         // For managing cancellation signals and deadlines across API boundaries and goroutines
+	"crypto/rand"     // For generating identity keys
+	"crypto/tls"      // For TLS configuration and certificates
+	"encoding/base64" // For encoding binary message data into JSON-safe strings
+	"encoding/binary" // For encoding/decoding length prefixes in stream communication
+	"encoding/json"   // For marshalling/unmarshalling data structures to/from JSON (used for C API communication)
+	"fmt"             // For formatted string creation and printing
+	"io"              // For input/output operations (e.g., reading from streams)
+	"log"             // For logging information, warnings, and errors
+	"net"             // For network-related errors and interfaces
+	"os"              // For interacting with the operating system (e.g., Stdout)
+	"path/filepath"   // For file path manipulations (e.g., saving/loading identity keys)
+	"strings"         // For string manipulations (e.g., trimming, splitting)
+	"sync"            // For synchronization primitives like Mutexes and RWMutexes to protect shared data
+	"time"            // For time-related functions (e.g., timeouts, timestamps)
+	"unsafe"          // For using Go pointers with C code (specifically C.free)
+
+	// Core libp2p libraries
+	libp2p "github.com/libp2p/go-libp2p"                          // Main libp2p package for creating a host
+	dht "github.com/libp2p/go-libp2p-kad-dht"                     // Kademlia DHT implementation for peer discovery and routing
+	"github.com/libp2p/go-libp2p/core/crypto"                     // Defines cryptographic primitives (keys, signatures)
+	"github.com/libp2p/go-libp2p/core/event"                      // Event bus for subscribing to libp2p events (connections, reachability changes)
+	"github.com/libp2p/go-libp2p/core/host"                       // Defines the main Host interface, representing a libp2p node
+	"github.com/libp2p/go-libp2p/core/network"                    // Defines network interfaces like Stream and Connection
+	"github.com/libp2p/go-libp2p/core/peer"                       // Defines Peer ID and AddrInfo types
+	"github.com/libp2p/go-libp2p/core/peerstore"                  // Defines the Peerstore interface for storing peer metadata (addresses, keys)
+	"github.com/libp2p/go-libp2p/core/routing"                    // Defines the Routing interface for peer routing (e.g., DHT)
+	rc "github.com/libp2p/go-libp2p/p2p/protocol/circuitv2/relay" // Import for relay service options
+	autorelay "github.com/libp2p/go-libp2p/p2p/host/autorelay"	   // AutoRelay for automatic relay selection and usage
+	rcmgr "github.com/libp2p/go-libp2p/p2p/host/resource-manager"			  // Resource manager for controlling resource usage (connections, streams)
+
+	// transport protocols for libp2p
+	quic "github.com/libp2p/go-libp2p/p2p/transport/quic"                 // QUIC transport for peer-to-peer connections (e.g., for mobile devices)
+	"github.com/libp2p/go-libp2p/p2p/transport/tcp"                       // TCP transport for peer-to-peer connections (most common)
+	webrtc "github.com/libp2p/go-libp2p/p2p/transport/webrtc"             // WebRTC transport for peer-to-peer connections (e.g., for browsers or mobile devices)
+	ws "github.com/libp2p/go-libp2p/p2p/transport/websocket"              // WebSocket transport for peer-to-peer connections (e.g., for browsers)
+	webtransport "github.com/libp2p/go-libp2p/p2p/transport/webtransport" // WebTransport transport for peer-to-peer connections (e.g., for browsers)
+
+	// --- AutoTLS Imports ---
+	"github.com/caddyserver/certmagic"                // Automatic TLS certificate management (used by p2p-forge)
+	golog "github.com/ipfs/go-log/v2"                 // IPFS logging library for structured logging
+	p2pforge "github.com/ipshipyard/p2p-forge/client" // p2p-forge library for automatic TLS and domain management
+
+	// protobuf
+	pg "unailib/proto-go" // Generated Protobuf code for our message formats
+
+	"google.golang.org/protobuf/proto" // Core Protobuf library for marshalling/unmarshalling messages
+
+	// PubSub library
+	pubsub "github.com/libp2p/go-libp2p-pubsub" // GossipSub implementation for publish/subscribe messaging
+
+	// Multiaddr libraries (libp2p's addressing format)
+	ma "github.com/multiformats/go-multiaddr"        // Core multiaddr parsing and manipulation
+)
+
+// ChatProtocol defines the protocol ID string used for direct peer-to-peer messaging streams.
+// This ensures that both peers understand how to interpret the data on the stream.
+// const UnaiverseChatProtocol = "/unaiverse-chat-protocol/1.0.0"
+const UnaiverseChatProtocol = "/unaiverse/chat/1.0.0"
+const UnaiverseUserAgent = "go-libp2p/example/autotls"
+const DisconnectionGracePeriod = 10 * time.Second
+
+// ExtendedPeerInfo holds information about a connected peer.
+type ExtendedPeerInfo struct {
+	ID          peer.ID        `json:"id"`           // the Peer ID of the connected peer.
+	Addrs       []ma.Multiaddr `json:"addrs"`        // the Multiaddr(s) associated with the peer.
+	ConnectedAt time.Time      `json:"connected_at"` // Timestamp when the connection was established.
+	Direction   string         `json:"direction"`    // Direction of the connection: "inbound" or "outbound".
+	Misc        int            `json:"misc"`         // Misc information (integer), custom usage
+	Relayed     bool           `json:"relayed"`      // Currently unused (but used in JS)
+}
+
+// RendezvousState holds the discovered peers from a rendezvous topic,
+// along with metadata about the freshness of the data.
+type RendezvousState struct {
+	Peers       map[peer.ID]ExtendedPeerInfo `json:"peers"`
+	UpdateCount int64                        `json:"update_count"`
+}
+
+// QueuedMessage represents a message received either directly or via PubSub.
+//
+// This lightweight version stores the binary payload in the `Data` field,
+// while the `From` field contains the Peer ID of the sender for security reasons.
+// It has to match with the 'sender' field in the ProtoBuf payload of the message.
+type QueuedMessage struct {
+	From peer.ID `json:"from"` // The VERIFIED peer ID of the sender from the network layer.
+	Data []byte  `json:"-"`    // The raw data payload (Protobuf encoded).
+}
+
+// MessageStore holds the QueuedMessages for each channel in separate FIFO queues.
+// It has a maximum number of channels and a maximum queue length per channel.
+type MessageStore struct {
+	mu                sync.Mutex            // protects the message store from concurrent access.
+	messagesByChannel map[string]*list.List // stores a FIFO queue of messages for each channel
+}
+
+// NodeConfig contains the parameters to initialize a node
+type NodeConfig struct {
+    IdentityDir     string   `json:"identity_dir"`
+    PredefinedPort  int      `json:"predefined_port"`
+    ListenIPs       []string `json:"listen_ips"`
+    
+    // Group Relay Logic
+    Relay struct {
+        EnableClient  bool `json:"enable_client"`
+        EnableService bool `json:"enable_service"`
+		WithBroadLimits bool `json:"with_broad_limits"`
+    } `json:"relay"`
+
+    // Group TLS Logic (Mutually exclusive logic becomes clear here)
+    TLS struct {
+        AutoTLS     bool   `json:"auto_tls"`
+        Domain      string `json:"domain"`
+        CertPath    string `json:"cert_path"`
+        KeyPath     string `json:"key_path"`
+    } `json:"tls"`
+
+    // explicit configuration for network environment
+    Network struct {
+        Isolated bool `json:"isolated"` // only allows connections with friendly peers
+		ForcePublic bool `json:"force_public"` // Replaces knowsIsPublic
+    } `json:"network"`
+
+	// Group DHT logic
+	DHT struct {
+		Enabled bool `json:"enabled"`
+		Keep bool	`json:"keep"` // to keep it running after init
+    } `json:"dht"`
+}
+
+// CreateNodeResponse defines the structure of our success message.
+type CreateNodeResponse struct {
+	Addresses []string `json:"addresses"`
+	IsPublic  bool     `json:"isPublic"`
+}
+
+// NodeInstance holds ALL state for a single libp2p node.
+type NodeInstance struct {
+	// Core Components
+	host         host.Host
+	pubsub       *pubsub.PubSub
+	dht          *dht.IpfsDHT
+	ctx          context.Context
+	cancel       context.CancelFunc
+	certManager  *p2pforge.P2PForgeCertMgr
+	messageStore *MessageStore
+
+	// Address Cache
+    addrMutex  sync.RWMutex
+    localAddrs []ma.Multiaddr
+
+	// Static relay
+	privateRelay *autorelay.AutoRelay
+	// privateRelayAddrs []ma.Multiaddr
+
+	// PubSub State
+	pubsubMutex   sync.RWMutex
+	topics        map[string]*pubsub.Topic
+	subscriptions map[string]*pubsub.Subscription
+
+	// Peer State
+	peersMutex     sync.RWMutex
+	friendlyPeers map[peer.ID]ExtendedPeerInfo
+
+	// Stream State
+	streamsMutex          sync.Mutex
+	persistentChatStreams map[peer.ID]network.Stream
+
+	// Disconnection Grace Period State
+    disconnectionMutex  sync.Mutex
+    disconnectionTimers map[peer.ID]context.CancelFunc
+
+	// Rendezvous State
+	rendezvousMutex sync.RWMutex
+	rendezvousState *RendezvousState
+
+	// a copy of its own index for logging
+	instanceIndex int
+}
+
+// --- Create a package-level logger ---
+var logger = golog.Logger("unailib")
+
+// --- Multi-Instance State Management ---
+var (
+	// Set the libp2p configuration parameters.
+	maxInstances       int
+	maxChannelQueueLen int
+	maxUniqueChannels  int
+	MaxMessageSize     uint32
+
+	// A single slice to hold all our instances.
+	allInstances []*NodeInstance
+	// A SINGLE mutex to protect the allInstances slice itself (during create/close).
+	globalInstanceMutex sync.RWMutex
+)
+
+// --- Helper Functions ---
+// jsonErrorResponse creates a JSON string representing an error state.
+// It takes a base message and an optional error, formats them, escapes the message
+// for JSON embedding, and returns a C string pointer (`*C.char`).
+// The caller (usually C/Python) is responsible for freeing this C string using FreeString.
+func jsonErrorResponse(
+	message string,
+	err error,
+) *C.char {
+
+	errMsg := message
+	if err != nil {
+		errMsg = fmt.Sprintf("%s: %s", message, err.Error())
+	}
+	logger.Errorf("[GO] ❌ Error: %s", errMsg)
+	// Ensure error messages are escaped properly for JSON embedding
+	escapedErrMsg := escapeStringForJSON(errMsg)
+	// Format into a standard {"state": "Error", "message": "..."} JSON structure.
+	jsonError := fmt.Sprintf(`{"state":"Error","message":"%s"}`, escapedErrMsg)
+	// Convert the Go string to a C string (allocates memory in C heap).
+	return C.CString(jsonError)
+}
+
+// jsonSuccessResponse creates a JSON string representing a success state.
+// It takes an arbitrary Go object (`message`), marshals it into JSON, wraps it
+// in a standard {"state": "Success", "message": {...}} structure, and returns
+// a C string pointer (`*C.char`).
+// The caller (usually C/Python) is responsible for freeing this C string using FreeString.
+func jsonSuccessResponse(
+	message interface{},
+) *C.char {
+
+	// Marshal the provided Go data structure into JSON bytes.
+	jsonData, err := json.Marshal(message)
+	if err != nil {
+		// If marshalling fails, return a JSON error response instead.
+		return jsonErrorResponse("Failed to marshal success response", err)
+	}
+	// Format into the standard success structure.
+	jsonSuccess := fmt.Sprintf(`{"state":"Success","message":%s}`, string(jsonData))
+	// Convert the Go string to a C string (allocates memory in C heap).
+	return C.CString(jsonSuccess)
+}
+
+// escapeStringForJSON performs basic escaping of characters (like double quotes and backslashes)
+// within a string to ensure it's safe to embed within a JSON string value.
+// It uses Go's standard JSON encoder for robust escaping.
+func escapeStringForJSON(
+	s string,
+) string {
+
+	var buf bytes.Buffer
+	// Encode the string using Go's JSON encoder, which handles escaping.
+	json.NewEncoder(&buf).Encode(s)
+	// The encoder adds surrounding quotes and a trailing newline, which we remove.
+	res := buf.String()
+	// Check bounds before slicing to avoid panic.
+	if len(res) > 2 && res[0] == '"' && res[len(res)-2] == '"' {
+		return res[1 : len(res)-2] // Trim surrounding quotes and newline
+	}
+	// Fallback if encoding behaves unexpectedly (e.g., empty string).
+	return s
+}
+
+// getInstance is a new helper to safely retrieve a node instance.
+// It handles bounds checking, nil checks, and locking.
+func getInstance(instanceIndex int) (*NodeInstance, error) {
+	if instanceIndex < 0 || instanceIndex >= maxInstances {
+		return nil, fmt.Errorf("invalid instance index: %d. Must be between 0 and %d", instanceIndex, maxInstances-1)
+	}
+
+	// Use a Read Lock, which is fast and allows concurrent reads.
+	globalInstanceMutex.RLock()
+	instance := allInstances[instanceIndex]
+	globalInstanceMutex.RUnlock()
+
+	if instance == nil {
+		return nil, fmt.Errorf("instance %d is not initialized or has been closed", instanceIndex)
+	}
+
+	// Check host as a proxy for *full* initialization,
+	// as it's set late in CreateNode
+	if instance.host == nil {
+		return nil, fmt.Errorf("instance %d is not fully initialized (host is nil)", instanceIndex)
+	}
+
+	return instance, nil
+}
+
+// newMessageStore initializes a new MessageStore.
+func newMessageStore() *MessageStore {
+	return &MessageStore{
+		messagesByChannel: make(map[string]*list.List),
+	}
+}
+
+func loadOrCreateIdentity(keyPath string) (crypto.PrivKey, error) {
+	// Check if key file already exists.
+	if _, err := os.Stat(keyPath); err == nil {
+		// Key file exists, read and unmarshal it.
+		bytes, err := os.ReadFile(keyPath)
+		if err != nil {
+			return nil, fmt.Errorf("failed to read existing key file: %w", err)
+		}
+		// load the key
+		privKey, err := crypto.UnmarshalPrivateKey(bytes)
+		if err != nil {
+			return nil, fmt.Errorf("failed to unmarshal corrupt private key: %w", err)
+		}
+		return privKey, nil
+
+	} else if os.IsNotExist(err) {
+		// Key file does not exist, generate a new one.
+		logger.Infof("[GO] 💎 Generating new persistent peer identity in %s\n", keyPath)
+		privKey, _, err := crypto.GenerateEd25519Key(rand.Reader)
+		if err != nil {
+			return nil, fmt.Errorf("failed to generate new key: %w", err)
+		}
+
+		// Marshal the new key to bytes.
+		bytes, err := crypto.MarshalPrivateKey(privKey)
+		if err != nil {
+			return nil, fmt.Errorf("failed to marshal new private key: %w", err)
+		}
+
+		// Write the new key to a file.
+		if err := os.WriteFile(keyPath, bytes, 0400); err != nil {
+			return nil, fmt.Errorf("failed to write new key file: %w", err)
+		}
+		return privKey, nil
+
+	} else {
+		// Another error occurred (e.g., permissions).
+		return nil, fmt.Errorf("failed to stat key file: %w", err)
+	}
+}
+
+func getListenAddrs(ips []string, tcpPort int, tlsMode string) ([]ma.Multiaddr, error) {
+	if len(ips) == 0 {
+        ips = []string{"0.0.0.0"}
+    }
+
+	var listenAddrs []ma.Multiaddr
+	quicPort := 0
+	webtransPort := 0
+	webrtcPort := 0
+	if tcpPort != 0 {
+		quicPort = tcpPort + 1
+		webtransPort = tcpPort + 2
+		webrtcPort = tcpPort +3
+	}
+
+	// --- Create Multiaddrs for both protocols from the single IP list ---
+	for _, ip := range ips {
+		// TCP
+		tcpMaddr, _ := ma.NewMultiaddr(fmt.Sprintf("/ip4/%s/tcp/%d", ip, tcpPort))
+		// QUIC
+		quicMaddr, _ := ma.NewMultiaddr(fmt.Sprintf("/ip4/%s/udp/%d/quic-v1", ip, quicPort))
+		// WebTransport
+		webtransMaddr, _ := ma.NewMultiaddr(fmt.Sprintf("/ip4/%s/udp/%d/quic-v1/webtransport", ip, webtransPort))
+		// WebRTC Direct
+		webrtcMaddr, _ := ma.NewMultiaddr(fmt.Sprintf("/ip4/%s/udp/%d/webrtc-direct", ip, webrtcPort))
+
+		listenAddrs = append(listenAddrs, tcpMaddr, quicMaddr, webtransMaddr, webrtcMaddr)
+
+		switch tlsMode {
+		case "autotls":
+			// This is the special multiaddr that triggers AutoTLS
+			wssMaddr, _ := ma.NewMultiaddr(fmt.Sprintf("/ip4/%s/tcp/%d/tls/sni/*.%s/ws", ip, tcpPort, p2pforge.DefaultForgeDomain))
+			listenAddrs = append(listenAddrs, wssMaddr)
+		case "domain":
+			// This is the standard secure WebSocket address with provided domain
+			wssMaddr, _ := ma.NewMultiaddr(fmt.Sprintf("/ip4/%s/tcp/%d/tls/ws", ip, tcpPort))
+			listenAddrs = append(listenAddrs, wssMaddr)
+		default:
+			// Fallback to a standard, non-secure WebSocket address
+			wsMaddr, _ := ma.NewMultiaddr(fmt.Sprintf("/ip4/%s/tcp/%d/ws", ip, tcpPort))
+			listenAddrs = append(listenAddrs, wsMaddr)
+		}
+	}
+
+	logger.Debugf("[GO] 🔧 Prepared Listen Addresses: %v\n", listenAddrs)
+
+	return listenAddrs, nil
+}
+
+func setupPubSub(ni *NodeInstance) error {
+	psOptions := []pubsub.Option{
+		// pubsub.WithFloodPublish(true),
+		pubsub.WithMaxMessageSize(int(MaxMessageSize)),
+	}
+	ps, err := pubsub.NewGossipSub(ni.ctx, ni.host, psOptions...)
+	if err != nil {
+		return err
+	}
+	ni.pubsub = ps // Set the pubsub field on the instance
+	return nil
+}
+
+func setupNotifiers(ni *NodeInstance) {
+	ni.host.Network().Notify(&network.NotifyBundle{
+		ConnectedF: func(_ network.Network, conn network.Conn) {
+			remotePeerID := conn.RemotePeer()
+			logger.Debugf("[GO] 🔔 Instance %d: Event - Connected to %s (Direction: %s)\n", ni.instanceIndex, remotePeerID, conn.Stat().Direction)
+			// --- Abort Graceful Disconnect if active ---
+            ni.disconnectionMutex.Lock()
+            if cancelTimer, exists := ni.disconnectionTimers[remotePeerID]; exists {
+                cancelTimer() // Stop the cleanup timer
+                delete(ni.disconnectionTimers, remotePeerID)
+                logger.Debugf("[GO] ♻️ Instance %d: Peer %s reconnected within grace period. Cleanup aborted.\n", ni.instanceIndex, remotePeerID)
+            }
+            ni.disconnectionMutex.Unlock()
+		},
+		DisconnectedF: func(_ network.Network, conn network.Conn) {
+			remotePeerID := conn.RemotePeer()
+			logger.Debugf("[GO] 🔔 Instance %d: Event - Disconnected from %s\n", ni.instanceIndex, remotePeerID)
+
+			// Get the host for this instance to query its network state.
+			if ni.host == nil {
+				// This shouldn't happen if the notifier is active, but a safe check.
+				logger.Warnf("[GO] ⚠️ Instance %d: DisconnectedF: Host is nil, cannot perform connection check.\n", ni.instanceIndex)
+				return
+			}
+
+			// Check if this is the LAST connection to this peer
+			if len(ni.host.Network().ConnsToPeer(remotePeerID)) == 0 {
+				// If it's a friendlyPeer, wait for the grace period, otherwise close immediately
+				ni.peersMutex.RLock()
+				_, isFriendly := ni.friendlyPeers[remotePeerID]
+				ni.peersMutex.RUnlock()
+
+				if isFriendly {
+					logger.Debugf("[GO] ⏳ Instance %d: Last connection to %s closed. Starting %v grace period timer...\n", ni.instanceIndex, remotePeerID, DisconnectionGracePeriod)
+
+					// We create a context that we can cancel if they reconnect
+					ctx, cancelTimer := context.WithCancel(context.Background())
+					
+					ni.disconnectionMutex.Lock()
+					// If a timer already exists (rare race condition), cancel the old one first
+					if oldCancel, exists := ni.disconnectionTimers[remotePeerID]; exists {
+						oldCancel()
+					}
+					ni.disconnectionTimers[remotePeerID] = cancelTimer
+					ni.disconnectionMutex.Unlock()
+
+					// Run cleanup in a goroutine
+					go func() {
+						select {
+						case <-time.After(DisconnectionGracePeriod):
+							// Timer expired! Proceed to cleanup.
+						case <-ctx.Done():
+							// Context cancelled (user reconnected). Stop here.
+							return
+						case <-ni.ctx.Done():
+							// Node is shutting down. Stop here.
+							return
+						}
+
+						// --- Timer Expired: Execute Cleanup ---
+						// Remove from timer map
+						ni.disconnectionMutex.Lock()
+						// Double-check: did we get cancelled while waiting for lock?
+						if ctx.Err() != nil {
+							ni.disconnectionMutex.Unlock()
+							return
+						}
+						delete(ni.disconnectionTimers, remotePeerID)
+						ni.disconnectionMutex.Unlock()
+
+						// Final Safety Check: Are they actually connected now?
+						// (Handles race where they reconnect exactly when timer fires)
+						if len(ni.host.Network().ConnsToPeer(remotePeerID)) > 0 {
+							logger.Debugf("[GO] ⚠️ Instance %d: Grace period expired for %s, but peer is connected again. Skipping cleanup.\n", ni.instanceIndex, remotePeerID)
+							return
+						}
+
+						logger.Debugf("[GO] 🗑️ Instance %d: Grace period ended for %s. Removing peer data.\n", ni.instanceIndex, remotePeerID)
+
+						// 3. Clean up friendlyPeers
+						ni.peersMutex.Lock()
+						if _, exists := ni.friendlyPeers[remotePeerID]; exists {
+							delete(ni.friendlyPeers, remotePeerID)
+							logger.Debugf("[GO]   Instance %d: Removed %s from friendlyPeers.\n", ni.instanceIndex, remotePeerID)
+						}
+						ni.peersMutex.Unlock()
+
+						// 4. Clean up persistent streams
+						ni.streamsMutex.Lock()
+						if stream, ok := ni.persistentChatStreams[remotePeerID]; ok {
+							logger.Debugf("[GO]   Instance %d: Cleaning up persistent stream for %s.\n", ni.instanceIndex, remotePeerID)
+							_ = stream.Close() 
+							delete(ni.persistentChatStreams, remotePeerID)
+						}
+						ni.streamsMutex.Unlock()
+
+					}()
+				} else {
+					logger.Debugf("[GO]   Instance %d: Last connection to %s closed. Removing from tracked peers.\n", ni.instanceIndex, remotePeerID)
+
+					// Also clean up persistent stream if one existed for this peer
+					ni.streamsMutex.Lock()
+					if stream, ok := ni.persistentChatStreams[remotePeerID]; ok {
+						logger.Debugf("[GO]   Instance %d: Cleaning up persistent stream for disconnected peer %s via DisconnectedF notifier.\n", ni.instanceIndex, remotePeerID)
+						_ = stream.Close() // Attempt graceful close
+						delete(ni.persistentChatStreams, remotePeerID)
+					}
+					ni.streamsMutex.Unlock()
+					}
+			} else {
+				logger.Debugf("[GO]   Instance %d: DisconnectedF: Still have %d active connections to %s, not removing.\n", ni.instanceIndex, len(ni.host.Network().ConnsToPeer(remotePeerID)), remotePeerID)
+			}
+		},
+	})
+}
+
+// enforceProtocolCompliance ensures that any connected peer supports the required chat protocol.
+// If a peer finishes identification but lacks the protocol, they are immediately disconnected.
+func enforceProtocolCompliance(ni *NodeInstance) {
+	// 1. Subscribe to the identification completed event
+	sub, err := ni.host.EventBus().Subscribe(new(event.EvtPeerIdentificationCompleted))
+	if err != nil {
+		logger.Errorf("[GO] ❌ Instance %d: Failed to subscribe to identification events: %v", ni.instanceIndex, err)
+		return
+	}
+
+	logger.Infof("[GO] 🛡️ Instance %d: Strict Isolation ENABLED. Monitoring for non-compliant peers.", ni.instanceIndex)
+
+	go func() {
+		defer sub.Close()
+		for {
+			select {
+			case <-ni.ctx.Done():
+				return
+			case evt, ok := <-sub.Out():
+				if !ok {
+					return
+				}
+				idEvt := evt.(event.EvtPeerIdentificationCompleted)
+
+				// Skip check for self
+				if idEvt.Peer == ni.host.ID() {
+					continue
+				}
+
+				isCompliant := false
+				for _, proto := range idEvt.Protocols {
+					if string(proto) == UnaiverseChatProtocol {
+						isCompliant = true
+						break
+					}
+				}
+
+				// 4. Action: Disconnect if not compliant
+				if !isCompliant {
+					logger.Warnf("[GO] 🚫 Instance %d: Kicking peer %s. (Reason: Protocol Mismatch).", ni.instanceIndex, idEvt.Peer)
+					// Disconnect
+					ni.host.Network().ClosePeer(idEvt.Peer)
+					// Optional: Clean from peerstore to free memory immediately
+					ni.host.Peerstore().RemovePeer(idEvt.Peer)
+				} else {
+					logger.Debugf("[GO] ✅ Instance %d: Peer %s verified compliant.", ni.instanceIndex, idEvt.Peer)
+				}
+			}
+		}
+	}()
+}
+
+// handleAddressUpdateEvents listens for libp2p address changes and updates the local cache.
+func handleAddressUpdateEvents(ni *NodeInstance, sub event.Subscription) {
+	defer sub.Close()
+
+	// Initialize cache immediately with current state to avoid race conditions at startup
+	ni.addrMutex.Lock()
+	ni.localAddrs = ni.host.Addrs()
+	ni.addrMutex.Unlock()
+
+	for {
+		select {
+		case <-ni.ctx.Done():
+			return
+		case _, ok := <-sub.Out():
+			if !ok {
+				return
+			}
+			// We only use the event as a trigger but we take the addresses from the Host
+			allAddresses := ni.host.Addrs()
+			ni.addrMutex.Lock()
+        	ni.localAddrs = allAddresses
+            ni.addrMutex.Unlock()
+            
+            // Log addresses to verify
+            addrsStr := make([]string, len(allAddresses))
+            for i, a := range allAddresses {
+                addrsStr[i] = a.String()
+            }
+			logger.Infof("[GO] 🔄 Instance %d: Updated local addresses (updating cache). Addrs: %v", ni.instanceIndex, addrsStr)
+		}
+	}
+}
+
+// Helper to filter peers for PeerSource
+func (ni *NodeInstance) isSuitableForPeerSource(pid peer.ID) bool {
+	ps := ni.host.Peerstore()
+
+	// 1. Check for Relay Hop Protocol (NEEDED)
+	protocols, err := ps.GetProtocols(pid)
+	if err != nil {
+		return false
+	}
+	isRelay := false
+	for _, proto := range protocols {
+		if proto == "/libp2p/circuit/relay/0.2.0/hop" {
+			isRelay = true
+			break
+		}
+	}
+	if !isRelay {
+		return false
+	}
+
+	// Only accept wss-enabled nodes as relay
+	addrs := ps.Addrs(pid)
+	isSuitable := false
+	for _, addr := range addrs {
+		_, err = addr.ValueForProtocol(ma.P_WS) 
+		if err == nil {
+			_, err := addr.ValueForProtocol(ma.P_TLS) 
+			if err == nil {
+				isSuitable = true
+			}
+		}
+	}
+
+	return isSuitable
+}
+
+// PeerSource acts as the peer discovery backend for AutoRelay.
+// It combines a local cache lookup (fast/free) with a DHT random walk (slow/expensive).
+func (ni *NodeInstance) PeerSource(ctx context.Context, numPeers int) <-chan peer.AddrInfo {
+	out := make(chan peer.AddrInfo)
+
+	go func() {
+		defer close(out)
+		
+		// Safety checks: Ensure host and DHT are fully initialized
+		if ni.host == nil || ni.dht == nil {
+			return
+		}
+
+		// Keep track of peers we've already sent in this batch
+		sentPeers := make(map[peer.ID]struct{})
+		peersFound := 0
+
+		// --- PHASE 1: Scavenge Local Peerstore ---
+		localPeers := ni.host.Peerstore().Peers()
+		for _, pid := range localPeers {
+			if peersFound >= numPeers {
+				return
+			}
+			if pid == ni.host.ID() {
+				continue
+			}
+
+			// Add it if it meets our criteria
+			if ni.isSuitableForPeerSource(pid) {
+				info := ni.host.Peerstore().PeerInfo(pid)
+				if len(info.Addrs) == 0 {
+					continue
+				}
+
+				select {
+				case out <- info:
+					sentPeers[pid] = struct{}{}
+					peersFound++
+				case <-ctx.Done():
+					return
+				}
+			}
+		}
+
+		// --- PHASE 2: DHT Random Walk ---
+		if peersFound < numPeers {
+			logger.Debugf("[GO] ⚠️ Instance %d: Local peerstore insufficient (%d/%d). Starting DHT walk...",
+				ni.instanceIndex, peersFound, numPeers)
+
+			for peersFound < numPeers {
+				randomKey := make([]byte, 32)
+				rand.Read(randomKey)
+				randomKeyStr := string(randomKey)
+
+				candidatePIDs, err := ni.dht.GetClosestPeers(ctx, randomKeyStr)
+				if err != nil {
+					select {
+					case <-ctx.Done():
+						return
+					case <-time.After(2 * time.Second):
+						continue
+					}
+				}
+
+				for _, pid := range candidatePIDs {
+					if peersFound >= numPeers {
+						return
+					}
+					if pid == ni.host.ID() {
+						continue
+					}
+					if _, alreadySent := sentPeers[pid]; alreadySent {
+						continue
+					}
+
+					info := ni.host.Peerstore().PeerInfo(pid)
+					if len(info.Addrs) > 0 {
+						select {
+						case out <- info:
+							sentPeers[pid] = struct{}{}
+							peersFound++
+						case <-ctx.Done():
+							return
+						}
+					}
+				}
+			}
+		}
+	}()
+
+	return out
+}
+
+// --- Core Logic Functions ---
+
+// storeReceivedMessage processes a raw message received either from a direct stream
+// or a PubSub topic. The sender peerID and the channel to store are retrieved in handleStream and readFromSubscription
+func storeReceivedMessage(
+	ni *NodeInstance,
+	from peer.ID,
+	channel string,
+	data []byte,
+) {
+	// Get the message store for this instance
+	store := ni.messageStore
+	if store == nil {
+		logger.Errorf("[GO] ❌ storeReceivedMessage: Message store not initialized for instance %d\n", ni.instanceIndex)
+		return // Cannot process message if store is nil
+	}
+
+	// Create the minimal message envelope.
+	newMessage := &QueuedMessage{
+		From: from,
+		Data: data,
+	}
+
+	// Lock the store mutex before accessing the shared maps.
+	store.mu.Lock()
+	defer store.mu.Unlock()
+
+	// Check if this channel already has a message list.
+	messageList, channelExists := store.messagesByChannel[channel]
+	if !channelExists {
+		// If the channel does not exist, check if we can create a new message queue.
+		if len(store.messagesByChannel) >= maxUniqueChannels {
+			logger.Warnf("[GO] 🗑️ Instance %d: Message store full. Discarding message for new channel '%s'.\n", ni.instanceIndex, channel)
+			return
+		}
+		messageList = list.New()
+		store.messagesByChannel[channel] = messageList
+		logger.Debugf("[GO] ✨ Instance %d: Created new channel queue '%s'. Total channels: %d\n", ni.instanceIndex, channel, len(store.messagesByChannel))
+	}
+
+	// If the channel already has a message list, check its length.
+	if messageList.Len() >= maxChannelQueueLen {
+		logger.Warnf("[GO] 🗑️ Instance %d: Queue for channel '%s' full. Discarding message.\n", ni.instanceIndex, channel)
+		return
+	}
+
+	messageList.PushBack(newMessage)
+	logger.Debugf("[GO] 📥 Instance %d: Queued message on channel '%s' from %s. New queue length: %d\n", ni.instanceIndex, channel, from, messageList.Len())
+}
+
+// readFromSubscription runs as a dedicated goroutine for each active PubSub subscription for a specific instance.
+// It continuously waits for new messages on the subscription's channel (`sub.Next(ctx)`),
+// routes them to `storeReceivedMessage`, and handles errors and context cancellation gracefully.
+// You need to provide the full Channel to uniquely identify the subscription.
+func readFromSubscription(
+	ni *NodeInstance,
+	sub *pubsub.Subscription,
+) {
+	// Get the topic string directly from the subscription object.
+	topic := sub.Topic()
+
+	if ni.ctx == nil || ni.host == nil {
+		logger.Errorf("[GO] ❌ readFromSubscription: Context or Host not initialized for instance %d. Exiting goroutine.\n", ni.instanceIndex)
+		return
+	}
+
+	logger.Infof("[GO] 👂 Instance %d: Started listener goroutine for topic: %s\n", ni.instanceIndex, topic)
+	defer logger.Infof("[GO] 👂 Instance %d: Exiting listener goroutine for topic: %s\n", ni.instanceIndex, topic) // Log when goroutine exits
+
+	for {
+		// Check if the main context has been cancelled (e.g., during node shutdown).
+		if ni.ctx.Err() != nil {
+			logger.Debugf("[GO] 👂 Instance %d: Context cancelled, stopping listener goroutine for topic: %s\n", ni.instanceIndex, topic)
+			return // Exit the goroutine.
+		}
+
+		// Wait for the next message from the subscription. This blocks until a message
+		// arrives, the context is cancelled, or an error occurs.
+		msg, err := sub.Next(ni.ctx)
+		if err != nil {
+			// Check for expected errors during shutdown or cancellation.
+			if err == context.Canceled || err == context.DeadlineExceeded || err == pubsub.ErrSubscriptionCancelled || ni.ctx.Err() != nil {
+				logger.Debugf("[GO] 👂 Instance %d: Subscription listener for topic '%s' stopping gracefully: %v\n", ni.instanceIndex, topic, err)
+				return // Exit goroutine cleanly.
+			}
+			// Handle EOF, which can sometimes occur. Treat it as a reason to stop.
+			if err == io.EOF {
+				logger.Debugf("[GO] 👂 Instance %d: Subscription listener for topic '%s' encountered EOF, stopping: %v\n", ni.instanceIndex, topic, err)
+				return // Exit goroutine.
+			}
+			// Log other errors but attempt to continue (they might be transient).
+			logger.Errorf("[GO] ❌ Instance %d: Error reading from subscription '%s': %v. Continuing...\n", ni.instanceIndex, topic, err)
+			// Pause briefly to avoid busy-looping on persistent errors.
+			time.Sleep(1 * time.Second)
+			continue // Continue the loop to try reading again.
+		}
+
+		logger.Infof("[GO] 📬 Instance %d (id: %s): Received new PubSub message on topic '%s' from %s\n", ni.instanceIndex, ni.host.ID().String(), topic, msg.GetFrom())
+
+		// Ignore messages published by the local node itself.
+		if msg.GetFrom() == ni.host.ID() {
+			continue // Skip processing self-sent messages.
+		}
+
+		// Handle Rendezvous or Standard Messages
+		if strings.HasSuffix(topic, ":rv") {
+			// This is a rendezvous update.
+			// 1. First, unmarshal the outer Protobuf message.
+			var protoMsg pg.Message
+			if err := proto.Unmarshal(msg.Data, &protoMsg); err != nil {
+				logger.Warnf("⚠️ Instance %d: Could not decode Protobuf message on topic '%s': %v\n", ni.instanceIndex, topic, err)
+				continue
+			}
+
+			// 2. The actual payload is a JSON string within the 'json_content' field.
+			jsonPayload := protoMsg.GetJsonContent()
+			if jsonPayload == "" {
+				logger.Warnf("⚠️ Instance %d: Rendezvous message on topic '%s' has empty JSON content.\n", ni.instanceIndex, topic)
+				continue
+			}
+
+			// 3. Now, unmarshal the inner JSON payload.
+			var updatePayload struct {
+				Peers       []ExtendedPeerInfo `json:"peers"`
+				UpdateCount int64              `json:"update_count"`
+			}
+			if err := json.Unmarshal([]byte(jsonPayload), &updatePayload); err != nil {
+				logger.Warnf("[GO] ⚠️ Instance %d: Could not decode rendezvous update payload on topic '%s': %v\n", ni.instanceIndex, topic, err)
+				continue // Skip this malformed message.
+			}
+
+			// 4. Create a new map from the decoded peer list.
+			newPeerMap := make(map[peer.ID]ExtendedPeerInfo)
+			for _, peerInfo := range updatePayload.Peers {
+				newPeerMap[peerInfo.ID] = peerInfo
+			}
+
+			// 5. Safely replace the old map with the new one.
+			ni.rendezvousMutex.Lock()
+			// If this is the first update for this instance, initialize the state struct.
+			if ni.rendezvousState == nil {
+				ni.rendezvousState = &RendezvousState{}
+			}
+			rendezvousState := ni.rendezvousState
+			rendezvousState.Peers = newPeerMap
+			rendezvousState.UpdateCount = updatePayload.UpdateCount
+			ni.rendezvousMutex.Unlock()
+
+			logger.Debugf("[GO] ✅ Instance %d: Updated rendezvous peers from topic '%s'. Found %d peers. Update count: %d.\n", ni.instanceIndex, topic, len(newPeerMap), updatePayload.UpdateCount)
+		} else {
+			// This is a standard message. Queue it as before.
+			logger.Debugf("[GO] 📝 Instance %d: Storing new pubsub message from topic '%s'.\n", ni.instanceIndex, topic)
+			storeReceivedMessage(ni, msg.GetFrom(), topic, msg.Data)
+		}
+	}
+}
+
+// handleStream reads from a direct message stream using the new framing protocol.
+// It expects the stream to start with a 4-byte length prefix, followed by a 1-byte channel name length,
+// the channel name itself, and finally the Protobuf-encoded payload.
+func handleStream(ni *NodeInstance, s network.Stream) {
+	senderPeerID := s.Conn().RemotePeer()
+	streamID := s.ID()
+	ni.peersMutex.Lock()
+	existingPeer, peerExists := ni.friendlyPeers[senderPeerID]
+
+	// 1. Gather fresh info (Addresses & Direction)
+	direction := "incoming"
+	if s.Stat().Direction == network.DirOutbound {
+		direction = "outgoing"
+	}
+	knownAddrs := ni.host.Peerstore().Addrs(senderPeerID)
+	if len(knownAddrs) == 0 {
+		knownAddrs = []ma.Multiaddr{s.Conn().RemoteMultiaddr()}
+	}
+
+	if !peerExists {
+		// CASE A: New Application Peer
+		ni.friendlyPeers[senderPeerID] = ExtendedPeerInfo{
+			ID:          senderPeerID,
+			Addrs:       knownAddrs,
+			ConnectedAt: time.Now(),
+			Direction:   direction,
+			Relayed:     false,
+		}
+		logger.Infof("[GO] ➕ Instance %d: Peer %s promoted to App Peer via Stream %s (Incoming).", ni.instanceIndex, senderPeerID, streamID)
+	} else {
+		// CASE B: Existing Peer - Update Addresses
+		// We keep ConnectedAt and Direction from the original session start.
+		existingPeer.Addrs = knownAddrs
+		ni.friendlyPeers[senderPeerID] = existingPeer
+		logger.Debugf("[GO] 🔄 Instance %d: Refreshed addresses for Peer %s via Stream %s.", ni.instanceIndex, senderPeerID, streamID)
+	}
+	ni.peersMutex.Unlock()
+	logger.Debugf("[GO] 📥 Instance %d: Accepted INCOMING stream %s from %s. Storing for duplex use.\n", ni.instanceIndex, streamID, senderPeerID)
+
+	// Store the newly accepted stream so we can use it to send messages back to this peer.
+	ni.streamsMutex.Lock()
+	ni.persistentChatStreams[senderPeerID] = s
+	ni.streamsMutex.Unlock()
+
+	// This defer block ensures cleanup happens when the stream is closed by either side.
+	defer func() {
+		logger.Debugf("[GO] 🧹 Instance %d: Stream %s with %s closed. Removing from map.\n", ni.instanceIndex, streamID, senderPeerID)
+		ni.streamsMutex.Lock()
+		if current, ok := ni.persistentChatStreams[senderPeerID]; ok && current == s {
+			delete(ni.persistentChatStreams, senderPeerID)
+		}
+		ni.streamsMutex.Unlock()
+		s.Close() // Ensure the stream is fully closed.
+	}()
+
+	for {
+		// Read 4-byte total length
+		var totalLen uint32
+		if err := binary.Read(s, binary.BigEndian, &totalLen); err != nil {
+			if err == io.EOF {
+				logger.Debugf("[GO] 🔌 Instance %d: Stream %s with %s closed (EOF).\n", ni.instanceIndex, streamID, senderPeerID)
+			} else if netErr, ok := err.(net.Error); ok && netErr.Timeout() {
+				logger.Warnf("[GO] ⏳ Instance %d: Timeout reading length from Stream %s (%s): %v\n", ni.instanceIndex, streamID, senderPeerID, err)
+			} else if errors.Is(err, network.ErrReset) {
+				logger.Warnf("[GO] ⚙️ Instance %d: Stream %s with %s reset.\n", ni.instanceIndex, streamID, senderPeerID)
+			} else {
+				logger.Errorf("[GO] ❌ Instance %d: Error reading length from Stream %s (%s): %v\n", ni.instanceIndex, streamID, senderPeerID, err)
+			}
+			return
+		}
+
+		if totalLen > MaxMessageSize {
+			logger.Errorf("[GO] ❌ Instance %d: Message len %d exceeds limit on Stream %s. Resetting.\n", ni.instanceIndex, totalLen, streamID)
+			s.Reset()
+			return
+		}
+
+		// Read Channel Length
+		var channelLen uint8
+		if err := binary.Read(s, binary.BigEndian, &channelLen); err != nil {
+			logger.Errorf("[GO] ❌ Instance %d: Error reading channel len from Stream %s: %v\n", ni.instanceIndex, streamID, err)
+			return
+		}
+
+		// Read Channel Name
+		channelBytes := make([]byte, channelLen)
+		if _, err := io.ReadFull(s, channelBytes); err != nil {
+			logger.Errorf("[GO] ❌ Instance %d: Error reading channel from Stream %s: %v\n", ni.instanceIndex, streamID, err)
+			return
+		}
+		channel := string(channelBytes)
+
+		// Read Payload
+		payloadLen := totalLen - uint32(channelLen) - 1
+		payload := make([]byte, payloadLen)
+		if _, err := io.ReadFull(s, payload); err != nil {
+			logger.Errorf("[GO] ❌ Instance %d: Error reading payload from Stream %s: %v\n", ni.instanceIndex, streamID, err)
+			return
+		}
+
+		logger.Infof("[GO] 📨 Instance %d: Received msg on channel '%s' via Stream %s from %s.\n", ni.instanceIndex, channel, streamID, senderPeerID)
+		storeReceivedMessage(ni, senderPeerID, channel, payload)
+	}
+}
+
+// setupDirectMessageHandler configures the libp2p host for a specific instance
+// to listen for incoming streams using the custom ChatProtocol.
+// When a peer opens a stream with this protocol ID, the provided handler function
+// is invoked to manage communication on that stream.
+func setupDirectMessageHandler(
+	ni *NodeInstance,
+) {
+	if ni.host == nil {
+		logger.Errorf("[GO] ❌ Instance %d: Cannot setup direct message handler: Host not initialized\n", ni.instanceIndex)
+		return
+	}
+
+	// Set a handler function for the UnaiverseChatProtocol. This function will be called
+	// automatically by libp2p whenever a new incoming stream for this protocol is accepted.
+	// Use a closure to capture the NodeInstance pointer.
+	ni.host.SetStreamHandler(UnaiverseChatProtocol, func(s network.Stream) {
+		handleStream(ni, s)
+	})
+}
+
+// This function constructs and writes a message using our new framing protocol for direct messages.
+// It takes a writer (e.g., a network stream), the channel name, and the payload data.
+// The message format is:
+// - 4-byte total length (including all the following parts)
+// - 1-byte channel name length
+// - channel name (as a UTF-8 string)
+// - payload (Protobuf-encoded data).
+func writeDirectMessageFrame(w io.Writer, channel string, payload []byte) error {
+	channelBytes := []byte(channel)
+	channelLen := uint8(len(channelBytes))
+
+	// Check if channel name is too long for our 1-byte length prefix.
+	if len(channelBytes) > 255 {
+		return fmt.Errorf("channel name exceeds 255 bytes limit: %s", channel)
+	}
+
+	// Total length = 1 (for channel len) + len(channel) + len(payload)
+	totalLength := uint32(1 + len(channelBytes) + len(payload))
+
+	// --- Add size check before writing ---
+	if totalLength > MaxMessageSize {
+		return fmt.Errorf("outgoing message size (%d) exceeds limit (%d)", totalLength, MaxMessageSize)
+	}
+
+	buf := new(bytes.Buffer)
+
+	// Write total length (4 bytes)
+	if err := binary.Write(buf, binary.BigEndian, totalLength); err != nil {
+		return fmt.Errorf("failed to write total length: %w", err)
+	}
+	// Write channel length (1 byte)
+	if err := binary.Write(buf, binary.BigEndian, channelLen); err != nil {
+		return fmt.Errorf("failed to write channel length: %w", err)
+	}
+	// Write channel name
+	if _, err := buf.Write(channelBytes); err != nil {
+		return fmt.Errorf("failed to write channel name: %w", err)
+	}
+	// Write payload
+	if _, err := buf.Write(payload); err != nil {
+		return fmt.Errorf("failed to write payload: %w", err)
+	}
+
+	// Write the entire frame to the stream.
+	if _, err := w.Write(buf.Bytes()); err != nil {
+		return fmt.Errorf("failed to write framed message to stream: %w", err)
+	}
+	return nil
+}
+
+// goGetNodeAddresses is the internal Go function that performs the core logic
+// of fetching and formatting node addresses.
+// It takes a pointer to a NodeInstance and a targetPID. If targetPID is empty (peer.ID("")),
+// it fetches addresses for the local node of the given instance.
+// It returns a slice of fully formatted multiaddress strings and an error if one occurs.
+func goGetNodeAddresses(
+	ni *NodeInstance,
+	targetPID peer.ID,
+) ([]string, error) {
+	if ni.host == nil {
+		errMsg := fmt.Sprintf("Instance %d: Host not initialized", ni.instanceIndex)
+		logger.Errorf("[GO] ❌ goGetNodeAddresses: %s\n", errMsg)
+		return nil, fmt.Errorf("%s", errMsg)
+	}
+
+	// Determine the actual Peer ID to resolve addresses for.
+	resolvedPID := targetPID
+	isThisNode := false
+	if targetPID == "" || targetPID == ni.host.ID() {
+		resolvedPID = ni.host.ID()
+		isThisNode = true
+	}
+
+	// --- 1. Gather all candidate addresses from the host and peerstore ---
+	var candidateAddrs []ma.Multiaddr
+	if isThisNode {
+		ni.addrMutex.RLock()
+		candidateAddrs = append(candidateAddrs, ni.localAddrs...)
+		// candidateAddrs = append(candidateAddrs, ni.privateRelayAddrs...)
+		ni.addrMutex.RUnlock()
+	} else {
+		// --- Remote Peer Addresses ---
+		ni.peersMutex.RLock()
+		if epi, exists := ni.friendlyPeers[resolvedPID]; exists {
+			candidateAddrs = append(candidateAddrs, epi.Addrs...)
+		}
+		ni.peersMutex.RUnlock()
+		candidateAddrs = append(candidateAddrs, ni.host.Peerstore().Addrs(resolvedPID)...)
+	}
+
+	// --- 2. Process and filter candidate addresses ---
+	addrSet := make(map[string]struct{})
+	for _, addr := range candidateAddrs {
+		// if addr == nil || manet.IsIPLoopback(addr) || manet.IsIPUnspecified(addr) {
+		// 	continue
+		// }
+
+		// Use the idiomatic `peer.SplitAddr` to check if the address already includes a Peer ID.
+		var finalAddr ma.Multiaddr
+		transportAddr, idInAddr := peer.SplitAddr(addr)
+		if transportAddr == nil {
+			continue
+		}
+
+		// handle cases for different transport protocols
+		if strings.HasPrefix(transportAddr.String(), "/p2p-circuit/") {
+			continue
+		}
+		if strings.Contains(transportAddr.String(), "*") {
+			continue
+		}
+
+		// handle cases based on presence and correctness of Peer ID in the address
+		switch {
+		case idInAddr == resolvedPID:
+			// Case A: The address is already perfect and has the correct Peer ID. Use it as is.
+			finalAddr = addr
+
+		case idInAddr == "":
+			// Case B: The address is missing a Peer ID. This is common for addresses from the
+			// peerstore and for relayed addresses like `/p2p/RELAY_ID/p2p-circuit`. We must append ours.
+			p2pComponent, _ := ma.NewMultiaddr(fmt.Sprintf("/p2p/%s", resolvedPID.String()))
+			finalAddr = addr.Encapsulate(p2pComponent)
+
+		case idInAddr != resolvedPID:
+			// Case C: The address has the WRONG Peer ID. This is stale or incorrect data. Discard it.
+			logger.Warnf("[GO] ⚠️ Instance %d: Discarding stale address for peer %s: %s\n", ni.instanceIndex, resolvedPID, addr)
+			continue
+		}
+		addrSet[finalAddr.String()] = struct{}{}
+	}
+
+	// --- 4. Convert the final set of unique addresses to a slice for returning. ---
+	result := make([]string, 0, len(addrSet))
+	for addr := range addrSet {
+		result = append(result, addr)
+	}
+
+	if len(result) == 0 {
+		logger.Warnf("[GO] ⚠️ goGetNodeAddresses: No suitable addresses found for peer %s.", resolvedPID)
+	}
+
+	return result, nil
+}
+
+// Close gracefully shuts down all components of this node instance.
+// This REPLACES the old `closeSingleInstance` function.
+func (ni *NodeInstance) Close() error {
+	logger.Infof("[GO] 🛑 Instance %d: Closing node...", ni.instanceIndex)
+
+	// --- Stop Cert Manager FIRST ---
+	if ni.certManager != nil {
+		logger.Debugf("[GO]   - Instance %d: Stopping AutoTLS cert manager...\n", ni.instanceIndex)
+		ni.certManager.Stop()
+	}
+
+	// --- Cancel Main Context ---
+	if ni.cancel != nil {
+		logger.Debugf("[GO]   - Instance %d: Cancelling main context...\n", ni.instanceIndex)
+		ni.cancel()
+	}
+
+	// Give goroutines time to react to context cancellation
+	time.Sleep(200 * time.Millisecond)
+
+	// --- Close DHT Client ---
+	if ni.dht != nil {
+		logger.Debugf("[GO]   - Instance %d: Closing DHT...\n", ni.instanceIndex)
+		if err := ni.dht.Close(); err != nil {
+			logger.Warnf("[GO] ⚠️ Instance %d: Error closing DHT: %v\n", ni.instanceIndex, err)
+		}
+		ni.dht = nil
+	}
+
+	// --- Close AutoRelay ---
+	if ni.privateRelay != nil {
+		logger.Debugf("[GO]   - Instance %d: Closing AutoRelay service...\n", ni.instanceIndex)
+		if err := ni.privateRelay.Close(); err != nil { //
+			logger.Warnf("[GO] ⚠️ Instance %d: Error closing AutoRelay: %v\n", ni.instanceIndex, err)
+		}
+		ni.privateRelay = nil
+	}
+
+	// --- Close Persistent Outgoing Streams ---
+	ni.streamsMutex.Lock()
+	if len(ni.persistentChatStreams) > 0 {
+		logger.Debugf("[GO]   - Instance %d: Closing %d persistent outgoing streams...\n", ni.instanceIndex, len(ni.persistentChatStreams))
+		for pid, stream := range ni.persistentChatStreams {
+			logger.Debugf("[GO]     - Instance %d: Closing stream to %s\n", ni.instanceIndex, pid)
+			_ = stream.Close() // Attempt graceful close
+		}
+	}
+	ni.persistentChatStreams = make(map[peer.ID]network.Stream) // Clear the map
+	ni.streamsMutex.Unlock()
+
+	// --- Clean Up PubSub State ---
+	ni.pubsubMutex.Lock()
+	if len(ni.subscriptions) > 0 {
+		logger.Debugf("[GO]   - Instance %d: Ensuring PubSub subscriptions (%d) are cancelled...\n", ni.instanceIndex, len(ni.subscriptions))
+		for channel, sub := range ni.subscriptions {
+			logger.Debugf("[GO]     - Instance %d: Cancelling subscription to topic: %s\n", ni.instanceIndex, channel)
+			sub.Cancel()
+		}
+	}
+	ni.subscriptions = make(map[string]*pubsub.Subscription) // Clear the map
+	ni.topics = make(map[string]*pubsub.Topic)               // Clear the map
+	ni.pubsubMutex.Unlock()
+
+	// --- Close Host Instance ---
+	var hostErr error
+	if ni.host != nil {
+		logger.Debugf("[GO]   - Instance %d: Closing host instance...\n", ni.instanceIndex)
+		hostErr = ni.host.Close()
+		if hostErr != nil {
+			logger.Warnf("[GO] ⚠️ %s (proceeding with cleanup)\n", hostErr)
+		} else {
+			logger.Debugf("[GO]   - Instance %d: Host closed successfully.\n", ni.instanceIndex)
+		}
+	}
+
+	// --- Clear Remaining State for this instance ---
+	ni.peersMutex.Lock()
+	ni.friendlyPeers = make(map[peer.ID]ExtendedPeerInfo) // Clear the map
+	ni.peersMutex.Unlock()
+
+	// Clear also the addresses
+	ni.addrMutex.Lock()
+	ni.localAddrs = nil
+	// ni.privateRelayAddrs = nil
+	ni.addrMutex.Unlock()
+
+	// Clear the MessageStore for this instance
+	if ni.messageStore != nil {
+		ni.messageStore.mu.Lock()
+		ni.messageStore.messagesByChannel = make(map[string]*list.List) // Clear the message store
+		ni.messageStore.mu.Unlock()
+	}
+	logger.Debugf("[GO]   - Instance %d: Cleared connected peers map and message buffer.\n", ni.instanceIndex)
+
+	// Clear the rendezvous state for this instance
+	ni.rendezvousMutex.Lock()
+	ni.rendezvousState = nil // Clear the state
+	ni.rendezvousMutex.Unlock()
+
+	// Explicitly cancel all running grace period timers so goroutines exit immediately.
+	ni.disconnectionMutex.Lock()
+	if len(ni.disconnectionTimers) > 0 {
+		logger.Debugf("[GO]   - Instance %d: Cancelling %d active disconnection timers...\n", ni.instanceIndex, len(ni.disconnectionTimers))
+		for _, cancelTimer := range ni.disconnectionTimers {
+			cancelTimer()
+		}
+	}
+	ni.disconnectionTimers = nil // Clear the map
+	ni.disconnectionMutex.Unlock()
+
+	// Nil out components to signify the instance is fully closed
+	ni.host = nil
+	ni.pubsub = nil
+	ni.ctx = nil
+	ni.cancel = nil
+	ni.certManager = nil
+	ni.messageStore = nil
+
+	if hostErr != nil {
+		return hostErr
+	}
+
+	logger.Infof("[GO] ✅ Instance %d: Node closed successfully.\n", ni.instanceIndex)
+	return nil
+}
+
+// --- Exported C Functions ---
+// These functions are callable from C (and thus Python). They act as the API boundary.
+
+// This function MUST be called once from Python before any other library function.
+//
+//export InitializeLibrary
+func InitializeLibrary(
+	maxInstancesC C.int,
+	maxUniqueChannelsC C.int,
+	maxChannelQueueLenC C.int,
+	maxMessageSizeC C.int,
+	logConfigJSONC *C.char,
+) {
+	// --- Configure Logging FIRST ---
+	log.SetFlags(log.LstdFlags | log.Lmicroseconds)
+	configStr := C.GoString(logConfigJSONC)
+	golog.SetAllLoggers(golog.LevelFatal)
+	if configStr != "" {
+		var logLevels map[string]string
+		if err := json.Unmarshal([]byte(configStr), &logLevels); err != nil {
+			log.Printf("[GO] ⚠️ Invalid log config JSON: %v. Using defaults.\n", err)
+		} else {
+			for logger, levelStr := range logLevels {
+				if err := golog.SetLogLevel(logger, levelStr); err != nil {
+					log.Printf("[GO] ⚠️ Failed to set log level for '%s': %v\n", logger, err)
+				}
+			}
+		}
+	}
+	
+	maxInstances = int(maxInstancesC)
+	maxUniqueChannels = int(maxUniqueChannelsC)
+	maxChannelQueueLen = int(maxChannelQueueLenC)
+	MaxMessageSize = uint32(maxMessageSizeC)
+
+	// Initialize the *single* global slice.
+	allInstances = make([]*NodeInstance, maxInstances)
+	logger.Infof("[GO] ✅ Go library initialized with MaxInstances=%d, MaxUniqueChannels=%d and MaxChannelQueueLen=%d\n", maxInstances, maxUniqueChannels, maxChannelQueueLen)
+}
+
+// CreateNode initializes and starts a new libp2p host (node) for a specific instance.
+// It configures the node based on the provided parameters (port, relay capabilities, UPnP).
+// Parameters:
+//   - instanceIndexC (C.int): The index for this node instance (0 to maxInstances-1).
+//   - predefinedPortC (C.int): The TCP port to listen on (0 for random).
+//   - enableRelayClientC (C.int): 1 if this node should enable relay communications (client mode)
+//   - enableRelayServiceC (C.int): 1 to set this node as a relay service (server mode),
+//   - knowsIsPublicC (C.int): 1 to assume public reachability, 0 otherwise (-> tries to assess it in any possible way).
+//   - maxConnectionsC (C.int): The maximum number of connections this node can maintain.
+//
+// Returns:
+//   - *C.char: A JSON string indicating success (with node addresses) or failure (with an error message).
+//     The structure is `{"state":"Success", "message": ["/ip4/.../p2p/...", ...]}` or `{"state":"Error", "message":"..."}`.
+//   - IMPORTANT: The caller (C/Python) MUST free the returned C string using the `FreeString` function
+//     exported by this library to avoid memory leaks. Returns NULL only on catastrophic failure before JSON creation.
+//
+//export CreateNode
+func CreateNode(
+	instanceIndexC C.int,
+	configJSONC *C.char,
+) (ret *C.char) {
+
+	instanceIndex := int(instanceIndexC)
+
+	if instanceIndex < 0 || instanceIndex >= maxInstances {
+		errMsg := fmt.Errorf("invalid instance index: %d. Must be between 0 and %d", instanceIndex, maxInstances-1)
+		return jsonErrorResponse("Invalid instance index", errMsg)
+	}
+
+	// --- Instance Creation and State Check ---
+	globalInstanceMutex.Lock()
+	if allInstances[instanceIndex] != nil {
+		globalInstanceMutex.Unlock()
+		msg := fmt.Sprintf("Instance %d is already initialized. Please call CloseNode first.", instanceIndex)
+		return jsonErrorResponse(msg, nil)
+	}
+
+	// --- Create the new instance object ---
+	ni := &NodeInstance{
+		instanceIndex:         instanceIndex,
+		topics:                make(map[string]*pubsub.Topic),
+		subscriptions:         make(map[string]*pubsub.Subscription),
+		friendlyPeers:         make(map[peer.ID]ExtendedPeerInfo),
+		persistentChatStreams: make(map[peer.ID]network.Stream),
+		disconnectionTimers:   make(map[peer.ID]context.CancelFunc),
+		messageStore:          newMessageStore(),
+	}
+	ni.ctx, ni.cancel = context.WithCancel(context.Background())
+	isPublic := false
+
+	// Store it in the global slice
+	allInstances[instanceIndex] = ni
+	globalInstanceMutex.Unlock()
+
+	logger.Infof("[GO] 🚀 Instance %d: Starting CreateNode...", instanceIndex)
+	// --- Centralized Cleanup on Failure ---
+	var success bool = false
+	defer func() {
+		if !success {
+			// If `success` is still false when CreateNode exits, an error
+			// must have occurred. We call Close() and remove the instance.
+			logger.Warnf("[GO] ⚠️ Instance %d: CreateNode failed, cleaning up...", instanceIndex)
+			ni.Close() // Call the new method!
+			globalInstanceMutex.Lock()
+			allInstances[instanceIndex] = nil // Remove it from the global list
+			globalInstanceMutex.Unlock()
+		}
+	}()
+
+	// 1. Parse Configuration
+	configJSON := C.GoString(configJSONC)
+	var cfg NodeConfig
+	if err := json.Unmarshal([]byte(configJSON), &cfg); err != nil {
+		return jsonErrorResponse("Invalid Configuration JSON", err)
+	}
+
+	// --- Sanity checks on the config ---
+	// If one of the three parameters for custom certificates is specified, all three are required.
+	if cfg.TLS.Domain != "" || cfg.TLS.CertPath != "" || cfg.TLS.KeyPath != "" {
+		if cfg.TLS.Domain == "" || cfg.TLS.CertPath == "" || cfg.TLS.KeyPath == "" {
+			return jsonErrorResponse(fmt.Sprintf("Instance %d: Missing at least one of 'Domain', 'CertPath' or 'KeyPath'.", instanceIndex), nil)
+		}
+	} // in the following, cfg.TLS.Domain != "" will be used as flag for useCustomTLS
+
+	// Having both customTLS and autoTLS is not allowed
+	if cfg.TLS.Domain != "" && cfg.TLS.AutoTLS {
+		return jsonErrorResponse(fmt.Sprintf("Instance %d: Cannot specify both a 'Domain' and 'AutoTLS'.", instanceIndex), nil)
+	}
+
+	// If we use AutoTLS we need the DHT on
+	if cfg.TLS.AutoTLS && !cfg.DHT.Enabled {
+		return jsonErrorResponse(fmt.Sprintf("Instance %d: Using TLS requires DHT 'Enabled'.", instanceIndex), nil)
+	}
+
+	// If we want RelayService we must be public (either forced or via AutoNat)
+	if cfg.Relay.EnableService {
+		if !cfg.Relay.EnableClient {
+			return jsonErrorResponse(fmt.Sprintf("Instance %d: Cannot set libp2p.DisableRelay() if we want to offer relay services.", instanceIndex), nil)
+		}
+		if !(cfg.DHT.Enabled || cfg.Network.ForcePublic) {
+			return jsonErrorResponse(fmt.Sprintf("Instance %d: A relay needs to be publicly reachable (forced or discovered).", instanceIndex), nil)
+		}
+	}
+
+	// If we want to keep dht it needs to be enabled
+	if cfg.DHT.Keep && !cfg.DHT.Enabled {
+		return jsonErrorResponse(fmt.Sprintf("Instance %d: Cannot set 'DHT.Keep' if DHT is not 'Enabled'.", instanceIndex), nil)
+	}
+
+	// --- Load or Create Persistent Identity ---
+	keyPath := filepath.Join(cfg.IdentityDir, "identity.key")
+	privKey, err := loadOrCreateIdentity(keyPath)
+	if err != nil {
+		return jsonErrorResponse(fmt.Sprintf("Instance %d: Failed to prepare identity", instanceIndex), err)
+	}
+
+	// --- AutoTLS Cert Manager Setup (if enabled) ---
+	var certManager *p2pforge.P2PForgeCertMgr
+	if cfg.TLS.AutoTLS {
+		logger.Debugf("[GO]   - Instance %d: AutoTLS is ENABLED. Setting up certificate manager...\n", instanceIndex)
+		certManager, err = p2pforge.NewP2PForgeCertMgr(
+			p2pforge.WithCAEndpoint(p2pforge.DefaultCAEndpoint),
+			p2pforge.WithCertificateStorage(&certmagic.FileStorage{Path: filepath.Join(cfg.IdentityDir, "p2p-forge-certs")}),
+			p2pforge.WithUserAgent(UnaiverseUserAgent),
+			p2pforge.WithRegistrationDelay(10*time.Second),
+		)
+		if err != nil {
+			return jsonErrorResponse(fmt.Sprintf("Instance %d: Failed to create AutoTLS cert manager", instanceIndex), err)
+		}
+		certManager.Start()
+		ni.certManager = certManager
+	}
+
+	// --- 4. Libp2p Options Assembly ---
+	tlsMode := "none"
+	if cfg.TLS.AutoTLS {
+		tlsMode = "autotls"
+	} else if cfg.TLS.Domain != "" {
+		tlsMode = "domain"
+	}
+	listenAddrs, err := getListenAddrs(cfg.ListenIPs, cfg.PredefinedPort, tlsMode)
+	if err != nil {
+		return jsonErrorResponse(fmt.Sprintf("Instance %d: Failed to create multiaddrs", instanceIndex), err)
+	}
+
+    // --- Configure Custom Resource Manager ---
+    scalingLimits := rcmgr.DefaultLimits
+    libp2p.SetDefaultServiceLimits(&scalingLimits)
+
+    // These apply per unique Peer ID.
+    scalingLimits.PeerBaseLimit.Conns = 64
+    scalingLimits.PeerBaseLimit.ConnsInbound = 64
+    scalingLimits.PeerBaseLimit.ConnsOutbound = 64
+
+    // Tweak System Limits
+    scalingLimits.SystemBaseLimit.Conns = 256
+    scalingLimits.SystemBaseLimit.ConnsInbound = 128
+    scalingLimits.SystemBaseLimit.ConnsOutbound = 128
+
+    // Compute the concrete limits
+    scaledLimits := scalingLimits.AutoScale()
+
+	// Raise the per-IP limits
+	customIP4Limits := []rcmgr.ConnLimitPerSubnet{
+        {
+            PrefixLength: 32,   // /32 means "one specific IP address"
+            ConnCount:    1024, // Allow 1024 conns from the same IP
+        },
+    }
+	customIP6Limits := []rcmgr.ConnLimitPerSubnet{
+        {
+            PrefixLength: 56,
+            ConnCount:    1024,
+        },
+    }
+
+    // Create the limiter and manager
+    limiter := rcmgr.NewFixedLimiter(scaledLimits)
+    rm, err := rcmgr.NewResourceManager(
+		limiter,
+		rcmgr.WithLimitPerSubnet(customIP4Limits, customIP6Limits),
+	)
+    if err != nil {
+        return jsonErrorResponse(fmt.Sprintf("Instance %d: Failed to create resource manager", instanceIndex), err)
+    }
+
+	options := []libp2p.Option{
+		libp2p.Identity(privKey),
+		libp2p.ListenAddrs(listenAddrs...),
+		libp2p.DefaultSecurity,
+		libp2p.DefaultMuxers,
+		libp2p.Transport(tcp.NewTCPTransport),
+		libp2p.ShareTCPListener(),
+		libp2p.Transport(quic.NewTransport),
+		libp2p.Transport(webtransport.New),
+		libp2p.Transport(webrtc.New),
+		libp2p.ResourceManager(rm),
+		libp2p.UserAgent(UnaiverseUserAgent),
+		libp2p.NATPortMap(),
+		libp2p.EnableHolePunching(),
+	}
+
+	// Add WebSocket transport, with or without TLS based on cert availability
+	if cfg.TLS.Domain != "" {
+		// We already have certificates, use them
+		logger.Debugf("[GO]   - Instance %d: Certificates provided, setting up secure WebSocket (WSS).\n", instanceIndex)
+		cert, err := tls.LoadX509KeyPair(cfg.TLS.CertPath, cfg.TLS.KeyPath)
+		if err != nil {
+			return jsonErrorResponse(fmt.Sprintf("Instance %d: Failed to load Custom TLS certificate and key", instanceIndex), err)
+		}
+		tlsConfig := &tls.Config{Certificates: []tls.Certificate{cert}}
+		// let's also create a custom address factory to ensure we always advertise the correct domain name
+		domainAddressFactory := func(addrs []ma.Multiaddr) []ma.Multiaddr {
+			// Replace the IP part of the WSS address with our domain.
+			result := make([]ma.Multiaddr, 0, len(addrs))
+			for _, addr := range addrs {
+				if strings.Contains(addr.String(), "/tls/ws") || strings.Contains(addr.String(), "/wss") {
+					// This is our WSS listener. Create the public /dns4 version.
+					portStr, err := addr.ValueForProtocol(ma.P_TCP)
+					if err != nil {
+						// Should not happen for a TCP/WS address, but safe fallback
+						result = append(result, addr)
+						continue
+					}
+					dnsAddr, _ := ma.NewMultiaddr(fmt.Sprintf("/dns4/%s/tcp/%s/tls/ws", cfg.TLS.Domain, portStr))
+					result = append(result, dnsAddr)
+				} else {
+					// Keep other addresses (like QUIC) as they are.
+					result = append(result, addr)
+				}
+			}
+			return result
+		}
+		options = append(options,
+			libp2p.Transport(ws.New, ws.WithTLSConfig(tlsConfig)),
+			libp2p.AddrsFactory(domainAddressFactory),
+		)
+		logger.Debugf("[GO]   - Instance %d: Loaded custom TLS certificate and key for WSS.\n", instanceIndex)
+	} else if cfg.TLS.AutoTLS {
+		// No certificates, create them automatically
+		options = append(options,
+			libp2p.Transport(ws.New, ws.WithTLSConfig(certManager.TLSConfig())),
+			libp2p.AddrsFactory(certManager.AddressFactory()),
+		)
+	} else {
+		// No certificates, use plain WS
+		logger.Debugf("[GO]   - Instance %d: No certificates found, setting up non-secure WebSocket.\n", instanceIndex)
+		options = append(options, libp2p.Transport(ws.New))
+	}
+
+	// Prepare discovering the bootstrap peers
+	if cfg.DHT.Enabled {
+		// Add any possible option to be publicly reachable
+		discoveryOpts := []libp2p.Option{
+			libp2p.EnableAutoNATv2(),
+			libp2p.Routing(func(h host.Host) (routing.PeerRouting, error) {
+				bootstrapAddrInfos := dht.GetDefaultBootstrapPeerAddrInfos()
+				dhtOptions := []dht.Option{
+					dht.Mode(dht.ModeClient),
+					dht.BootstrapPeers(bootstrapAddrInfos...),
+				}
+				var err error
+				ni.dht, err = dht.New(ni.ctx, h, dhtOptions...)
+				return ni.dht, err
+			}),
+		}
+		options = append(options, discoveryOpts...)
+		logger.Debugf("[GO]   - Instance %d: Trying to be publicly reachable.\n", instanceIndex)
+	}
+
+	// EnableRelay (the ability to *use* relays) is default, we can explicitly disable it if needed.
+	if !cfg.Relay.EnableClient {
+		// In this case we don't want to use the circuit-relay protocol.
+		options = append(options, libp2p.DisableRelay()) // Explicitly disable using relays.
+		logger.Debugf("[GO]   - Instance %d: Relay client is DISABLED.\n", instanceIndex)
+	} else {
+		// Configure Relay Service (ability to *be* a relay)
+		if cfg.Relay.EnableService {
+			resources := rc.DefaultResources() // open this to see the default resource limits
+			resources.MaxReservations = 1024			// default is 128
+			resources.MaxCircuits = 32					// default is 16
+			resources.BufferSize = 4096					// default is 2048
+			resources.MaxReservationsPerIP = 1024		// default is 8
+			resources.MaxReservationsPerASN = 1024		// default is 32
+			if cfg.Relay.WithBroadLimits {
+				// Enrich default limits
+				resources.Limit = nil // same as setting rc.WithInfiniteLimits()
+				logger.Debugf("[GO]   - Instance %d: Relay service is ENABLED with custom resource configuration (WithBroadLimits).\n", instanceIndex)
+			} else {
+				logger.Debugf("[GO]   - Instance %d: Relay service is ENABLED with default resource configuration.\n", instanceIndex)
+			}
+			// This single option enables the node to act as a relay for others.
+			options = append(options, libp2p.EnableRelayService(rc.WithResources(resources)), libp2p.EnableNATService())
+		} else {
+			// In this case we want to use relays but not offer the service to others.
+			// If we are exploiting the DHT we can start an AutoRelay with PeerSource
+			if cfg.DHT.Keep {
+				// Enable AutoRelay. This uses the services above (DHT, AutoNAT)
+				// to find relays and bind to one if we are private.
+				options = append(options, libp2p.EnableAutoRelayWithPeerSource(ni.PeerSource, autorelay.WithBootDelay(time.Second*10)))
+				logger.Debugf("[GO]   - Instance %d: AutoRelay client ENABLED.\n", instanceIndex)
+			}
+		}
+	}
+
+	if cfg.Network.ForcePublic {
+		// Force public reachability to test local relays
+		options = append(options, libp2p.ForceReachabilityPublic())
+	}
+
+	// Create the libp2p Host instance with the configured options for this instance.
+	host, err := libp2p.New(options...)
+	if err != nil {
+		return jsonErrorResponse(fmt.Sprintf("Instance %d: Failed to create host", instanceIndex), err)
+	}
+	ni.host = host
+	logger.Infof("[GO] ✅ Instance %d: Host created with ID: %s\n", instanceIndex, ni.host.ID())
+
+	if cfg.Network.Isolated {
+        // Turn on the "Protocol Police"
+        enforceProtocolCompliance(ni)
+    }
+
+	// --- Link Host to Cert Manager ---
+	if cfg.TLS.AutoTLS {
+		certManager.ProvideHost(ni.host)
+		logger.Debugf("[GO]   - Instance %d: Provided host to AutoTLS cert manager.\n", instanceIndex)
+	}
+
+	// --- Start Address Reporting & Caching ---
+	cacheSub, err := ni.host.EventBus().Subscribe(new(event.EvtLocalAddressesUpdated))
+	if err != nil {
+		return jsonErrorResponse(fmt.Sprintf("Instance %d: Failed to create address cache subscription", instanceIndex), err)
+	}
+	go handleAddressUpdateEvents(ni, cacheSub)
+	logger.Debugf("[GO] 🧠 Instance %d: Address cache background listener started.", instanceIndex)
+
+	if cfg.Network.ForcePublic {
+		isPublic = true
+		logger.Debugf("[GO] ⏳ Instance %d: ForcePublic is ON. Waiting for addresses to settle...", instanceIndex)
+		waitCtx, waitCancel := context.WithTimeout(ni.ctx, 5*time.Second)
+		defer waitCancel()
+
+		ticker := time.NewTicker(100 * time.Millisecond)
+		defer ticker.Stop()
+
+		AddressWaitLoop:
+		for {
+			select {
+			case <-waitCtx.Done():
+				logger.Warnf("[GO] ⚠️ Instance %d: Timed out waiting for addresses (proceeding anyway).", instanceIndex)
+				break AddressWaitLoop
+			case <-ticker.C:
+				// Check if the host has reported addresses yet
+				if len(ni.host.Addrs()) > 0 {
+					logger.Debugf("[GO] ✅ Instance %d: Addresses populated.", instanceIndex)
+					break AddressWaitLoop
+				}
+			}
+		}
+	} else {
+		// --- Wait for Reachability Update ---
+		// Subscribe to reachability events
+		reachSub, err := ni.host.EventBus().Subscribe(new(event.EvtLocalReachabilityChanged))
+		if err != nil {
+			return jsonErrorResponse("Failed to subscribe to reachability events", err)
+		}
+		defer reachSub.Close()
+
+		timeoutCtx, timeoutCancel := context.WithTimeout(ni.ctx, 30*time.Second)
+		defer timeoutCancel()
+		logger.Debugf("[GO] ⏳ Instance %d: Waiting for reachability update.", instanceIndex)
+		
+		WAIT_LOOP:
+		for {		
+			select {
+			case evt := <-reachSub.Out():
+				rEvt := evt.(event.EvtLocalReachabilityChanged)
+				if rEvt.Reachability == network.ReachabilityPublic {
+					logger.Debugf("[GO] 📶 Instance %d: Reachability -> PUBLIC", instanceIndex)
+					isPublic = true
+				} else {
+					isPublic = false
+				}
+				break WAIT_LOOP
+		
+			case <-timeoutCtx.Done():
+				logger.Warnf("[GO] ⚠️ Instance %d: Timeout. Proceeding with best effort. (Public: %t)", instanceIndex, isPublic)
+				break WAIT_LOOP
+
+			// 4. Node Shutdown
+			case <-ni.ctx.Done():
+				return jsonErrorResponse("Context cancelled during init", nil)
+			}
+		}
+	}
+
+	// --- PubSub Initialization ---
+	if err := setupPubSub(ni); err != nil {
+		return jsonErrorResponse(fmt.Sprintf("Instance %d: Failed to create PubSub", instanceIndex), err)
+	}
+	logger.Debugf("[GO] ✅ Instance %d: PubSub (GossipSub) initialized.\n", instanceIndex)
+
+	// --- Setup Notifiers and Handlers ---
+	setupNotifiers(ni)
+	logger.Debugf("[GO] 🔔 Instance %d: Registered network event notifier.\n", instanceIndex)
+
+	setupDirectMessageHandler(ni)
+	logger.Debugf("[GO] ✅ Instance %d: Direct message handler set up.\n", instanceIndex)
+
+	// --- Close DHT if needed ---
+	if !cfg.DHT.Keep {
+		if ni.dht != nil {
+			logger.Debugf("[GO]   - Instance %d: Closing DHT client...\n", instanceIndex)
+			if err := ni.dht.Close(); err != nil {
+				logger.Warnf("[GO] ⚠️ Instance %d: Error closing DHT: %v\n", instanceIndex, err)
+			}
+			ni.dht = nil
+		}
+	}
+
+	// --- Get Final Addresses ---
+	nodeAddresses, err := goGetNodeAddresses(ni, "")
+	if err != nil {
+		return jsonErrorResponse(
+			fmt.Sprintf("Instance %d: Failed to obtain node addresses after waiting for reachability", instanceIndex),
+			err,
+		)
+	}
+
+	// --- Build and return the new structured response ---
+	response := CreateNodeResponse{
+		Addresses: nodeAddresses,
+		IsPublic:  isPublic,
+	}
+
+	logger.Infof("[GO] 🌐 Instance %d: Node addresses: %v\n", instanceIndex, nodeAddresses)
+	reachabilityStatus := map[bool]string{true: "Public", false: "Private"}[isPublic]
+	logger.Infof("[GO] 🎉 Instance %d: Node creation complete. Reachability status: %s\n", instanceIndex, reachabilityStatus)
+	success = true // Mark success to avoid cleanup in defer.
+	return jsonSuccessResponse(response)
+}
+
+// ConnectTo attempts to establish a connection with a remote peer given its multiaddress for a specific instance.
+// Parameters:
+//   - instanceIndexC (C.int): The index of the node instance.
+//   - addrsJSONC (*C.char): Pointer to a JSON string containing the list of addresses that can be dialed.
+//
+// Returns:
+//   - *C.char: A JSON string indicating success (with peer AddrInfo of the winning connection) or failure (with an error message).
+//     Structure: `{"state":"Success", "message": {"ID": "...", "Addrs": ["...", ...]}}` or `{"state":"Error", "message":"..."}`.
+//   - IMPORTANT: The caller MUST free the returned C string using `FreeString`.
+//
+//export ConnectTo
+func ConnectTo(
+	instanceIndexC C.int,
+	addrsJSONC *C.char,
+) *C.char {
+
+	ni, err := getInstance(int(instanceIndexC))
+	if err != nil {
+		return jsonErrorResponse("Invalid instance", err)
+	}
+	
+	goAddrsJSON := C.GoString(addrsJSONC)
+	logger.Debugf("[GO] 📞 Instance %d: Attempting to connect to peer with addresses: %s\n", ni.instanceIndex, goAddrsJSON)
+
+	// --- Unmarshal Address List from JSON ---
+	var addrStrings []string
+	if err := json.Unmarshal([]byte(goAddrsJSON), &addrStrings); err != nil {
+		return jsonErrorResponse("Failed to parse addresses JSON", err)
+	}
+	if len(addrStrings) == 0 {
+		return jsonErrorResponse("Address list is empty", nil)
+	}
+
+	// --- Create AddrInfo from the list ---
+	addrInfo, err := peer.AddrInfoFromString(addrStrings[0])
+	if err != nil {
+		return jsonErrorResponse("Invalid first multiaddress in list", err)
+	}
+
+	// Add the rest of the addresses to the AddrInfo struct
+	for i := 1; i < len(addrStrings); i++ {
+		maddr, err := ma.NewMultiaddr(addrStrings[i])
+		if err != nil {
+			logger.Warnf("[GO] ⚠️ Instance %d: Skipping invalid multiaddress '%s' in list: %v\n", ni.instanceIndex, addrStrings[i], err)
+			continue
+		}
+		// You might want to add a check here to ensure subsequent addresses are for the same peer ID
+		addrInfo.Addrs = append(addrInfo.Addrs, maddr)
+	}
+
+	// Check if attempting to connect to the local node itself.
+	if addrInfo.ID == ni.host.ID() {
+		logger.Debugf("[GO] ℹ️ Instance %d: Attempting to connect to self (%s), skipping explicit connection.\n", ni.instanceIndex, addrInfo.ID)
+		// Connecting to self is usually not necessary or meaningful in libp2p.
+		// Return success, indicating the "connection" is implicitly present.
+		return jsonSuccessResponse(addrInfo) // Caller frees.
+	}
+
+	// --- 1. ESTABLISH CONNECTION ---
+	// Use a context with a timeout for the connection attempt to prevent blocking indefinitely.
+	connCtx, cancel := context.WithTimeout(ni.ctx, 30*time.Second) // 30-second timeout.
+	defer cancel()                                                      // Ensure context is cancelled eventually.
+
+	// Add the peer's address(es) to the local peerstore for this instance. This helps libp2p find the peer.
+	// ConnectedAddrTTL suggests the address is likely valid for a short time after connection.
+	// Use PermanentAddrTTL if the address is known to be stable.
+	ni.host.Peerstore().AddAddrs(addrInfo.ID, addrInfo.Addrs, peerstore.ConnectedAddrTTL)
+
+	// Initiate the connection attempt. libp2p will handle dialing and negotiation.
+	logger.Debugf("[GO]   - Instance %d: Attempting host.Connect to %s...\n", ni.instanceIndex, addrInfo.ID)
+	if err := ni.host.Connect(connCtx, *addrInfo); err != nil {
+		// Check if the error was due to the connection timeout.
+		if connCtx.Err() == context.DeadlineExceeded {
+			errMsg := fmt.Sprintf("Instance %d: Connection attempt to %s timed out after 30s", ni.instanceIndex, addrInfo.ID)
+			logger.Errorf("[GO] ❌ %s\n", errMsg)
+			return jsonErrorResponse(errMsg, nil) // Return specific timeout error (caller frees).
+		}
+		// Handle other connection errors.
+		errMsg := fmt.Sprintf("Instance %d: Failed to connect to peer %s", ni.instanceIndex, addrInfo.ID)
+		// Example: Check for specific common errors if needed
+		// if strings.Contains(err.Error(), "no route to host") { ... }
+		return jsonErrorResponse(errMsg, err) // Return generic connection error (caller frees).
+	}
+
+	// --- 2. FIND THE WINNING ADDRESS ---
+	// After a successful connection, query the host's network for active connections to the peer.
+	// This is where you find the 'winning' address.
+	conns := ni.host.Network().ConnsToPeer(addrInfo.ID)
+	var winningAddr string
+	if len(conns) > 0 {
+		winningAddr = fmt.Sprintf("%s/p2p/%s", conns[0].RemoteMultiaddr().String(), addrInfo.ID.String())
+		logger.Debugf("[GO] ✅ Instance %d: Successfully connected to peer %s via: %s\n", ni.instanceIndex, addrInfo.ID, winningAddr)
+	} else {
+		logger.Warnf("[GO] ⚠️ Instance %d: Connect succeeded for %s, but no active connection found immediately. It may be pending.\n", ni.instanceIndex, addrInfo.ID)
+	}
+
+	// Success: log the successful connection and return the response.
+	logger.Infof("[GO] ✅ Instance %d: Successfully initiated connection to multiaddress: %s\n", ni.instanceIndex, winningAddr)
+	winningAddrInfo, err := peer.AddrInfoFromString(winningAddr)
+	if err != nil {
+		return jsonErrorResponse("Invalid winner multiaddress.", err)
+	}
+	return jsonSuccessResponse(winningAddrInfo) // Caller frees.
+}
+
+// StartStaticRelay configures and starts the AutoRelay service using a specific
+// static relay (e.g., the subnetwork owner). This replaces manual reservation logic.
+//
+// Parameters:
+//   - instanceIndexC: The node instance.
+//   - relayAddrInfoJSONC: JSON string of the relay's AddrInfo (id + addrs).
+//
+//export StartStaticRelay
+func StartStaticRelay(
+	instanceIndexC C.int,
+	relayAddrInfoJSONC *C.char,
+) *C.char {
+
+	ni, err := getInstance(int(instanceIndexC))
+	if err != nil {
+		return jsonErrorResponse("Invalid instance", err)
+	}
+
+	// --- 1. Handle Switching Subnetworks ---
+	// If an AutoRelay service is already running, close it first.
+	if ni.privateRelay != nil {
+		logger.Debugf("[GO] 🔄 Instance %d: Switching Relay. Closing existing AutoRelay service...", ni.instanceIndex)
+		if err := ni.privateRelay.Close(); err != nil { //
+			logger.Warnf("[GO] ⚠️ Instance %d: Error closing old AutoRelay: %v", ni.instanceIndex, err)
+		}
+		ni.privateRelay = nil
+		// // Also clean up any existing relayed addresses
+		// ni.addrMutex.Lock()
+		// ni.privateRelayAddrs = nil
+		// ni.addrMutex.Unlock()
+		logger.Infof("[GO] 🔄 Instance %d: Previous AutoRelay service closed.", ni.instanceIndex)
+	}
+
+	// --- 2. Parse the New Relay's AddrInfo ---
+	relayInfoJSON := C.GoString(relayAddrInfoJSONC)
+	var relayInfo peer.AddrInfo
+	if err := json.Unmarshal([]byte(relayInfoJSON), &relayInfo); err != nil {
+		return jsonErrorResponse("Failed to parse relay AddrInfo JSON", err)
+	}
+
+	logger.Debugf("[GO] 🔗 Instance %d: Configuring Static AutoRelay with peer %s", ni.instanceIndex, relayInfo.ID)
+
+	// --- 3. Configure AutoRelay Options ---
+	opts := []autorelay.Option{
+		autorelay.WithStaticRelays([]peer.AddrInfo{relayInfo}),
+		autorelay.WithNumRelays(1), 
+		autorelay.WithBootDelay(0), 
+	}
+
+	// --- 4. Create the AutoRelay Service ---
+	// This initializes the service but might not start the background workers yet.
+	ar, err := autorelay.NewAutoRelay(ni.host, opts...) //
+	if err != nil {
+		return jsonErrorResponse("Failed to create AutoRelay service", err)
+	}
+
+	// --- 5. Start the Service ---
+	// This kicks off the background goroutines to connect and reserve slots.
+    // It returns immediately.
+	ar.Start() 
+
+	// Store the reference so we can close it later.
+	ni.privateRelay = ar
+
+	logger.Infof("[GO] ✅ Instance %d: Static AutoRelay service started. Target: %s", ni.instanceIndex, relayInfo.ID)
+
+	return jsonSuccessResponse("Static AutoRelay enabled")
+}
+
+// DisconnectFrom attempts to close any active connections to a specified peer
+// and removes the peer from the internally tracked list for a specific instance.
+// Parameters:
+//   - instanceIndexC (C.int): The index of the node instance.
+//   - peerIDC (*C.char): The Peer ID string of the peer to disconnect from.
+//
+// Returns:
+//   - *C.char: A JSON string indicating success or failure.
+//     Structure: `{"state":"Success", "message":"Disconnected from peer ..."}` or `{"state":"Error", "message":"..."}`.
+//   - IMPORTANT: The caller MUST free the returned C string using `FreeString`.
+//
+//export DisconnectFrom
+func DisconnectFrom(
+	instanceIndexC C.int,
+	peerIDC *C.char,
+) *C.char {
+
+	ni, err := getInstance(int(instanceIndexC))
+	if err != nil {
+		return jsonErrorResponse("Invalid instance", err)
+	}
+
+	goPeerID := C.GoString(peerIDC)
+	logger.Debugf("[GO] 🔌 Instance %d: Attempting to disconnect from peer: %s\n", ni.instanceIndex, goPeerID)
+
+	pid, err := peer.Decode(goPeerID)
+	if err != nil {
+		return jsonErrorResponse(
+			fmt.Sprintf("Instance %d: Failed to decode peer ID", ni.instanceIndex), err,
+		)
+	}
+
+	if pid == ni.host.ID() {
+		logger.Debugf("[GO] ℹ️ Instance %d: Attempting to disconnect from self (%s), skipping.\n", ni.instanceIndex, pid)
+		return jsonSuccessResponse("Cannot disconnect from self")
+	}
+
+	// --- Close Persistent Outgoing Stream (if exists) for this instance ---
+	ni.streamsMutex.Lock()
+	stream, exists := ni.persistentChatStreams[pid]
+	if exists {
+		logger.Debugf("[GO]   ↳ Instance %d: Closing persistent outgoing stream to %s\n", ni.instanceIndex, pid)
+		_ = stream.Close() // Attempt graceful close
+		delete(ni.persistentChatStreams, pid)
+	}
+	ni.streamsMutex.Unlock() // Unlock before potentially blocking network call
+
+	// --- Close Network Connections ---
+	conns := ni.host.Network().ConnsToPeer(pid)
+	closedNetworkConn := false
+	if len(conns) > 0 {
+		logger.Debugf("[GO]   - Instance %d: Closing %d active network connection(s) to peer %s...\n", ni.instanceIndex, len(conns), pid)
+		err = ni.host.Network().ClosePeer(pid) // This closes the underlying connection(s)
+		if err != nil {
+			logger.Warnf("[GO] ⚠️ Instance %d: Error closing network connection(s) to peer %s: %v (proceeding with cleanup)\n", ni.instanceIndex, pid, err)
+		} else {
+			logger.Debugf("[GO]   - Instance %d: Closed network connection(s) to peer: %s\n", ni.instanceIndex, pid)
+			closedNetworkConn = true
+		}
+	} else {
+		logger.Debugf("[GO] ℹ️ Instance %d: No active network connections found to peer %s.\n", ni.instanceIndex, pid)
+	}
+
+	// --- Remove from Tracking Map for this instance ---
+	ni.peersMutex.Lock()
+	delete(ni.friendlyPeers, pid)
+	ni.peersMutex.Unlock()
+
+	logMsg := fmt.Sprintf("Instance %d: Disconnected from peer %s", ni.instanceIndex, goPeerID)
+	if !exists && !closedNetworkConn && len(conns) == 0 {
+		logMsg = fmt.Sprintf("Instance %d: Disconnected from peer %s (not connected or tracked)", ni.instanceIndex, goPeerID)
+	}
+	logger.Infof("[GO] 🔌 %s\n", logMsg)
+
+	return jsonSuccessResponse(logMsg)
+}
+
+// GetConnectedPeers returns a list of peers currently tracked as connected for a specific instance.
+// Note: This relies on the internal `connectedPeersInstances` map which is updated during
+// connect/disconnect operations and incoming streams. It may optionally perform
+// a liveness check.
+// Parameters:
+//   - instanceIndexC (C.int): The index of the node instance.
+//
+// Returns:
+//   - *C.char: A JSON string containing a list of connected peers' information.
+//     Structure: `{"state":"Success", "message": [ExtendedPeerInfo, ...]}` or `{"state":"Error", "message":"..."}`.
+//     Each `ExtendedPeerInfo` object has `addr_info` (ID, Addrs), `connected_at`, `direction`, and `misc`.
+//   - IMPORTANT: The caller MUST free the returned C string using `FreeString`.
+//
+//export GetConnectedPeers
+func GetConnectedPeers(
+	instanceIndexC C.int,
+) *C.char {
+
+	ni, err := getInstance(int(instanceIndexC))
+	if err != nil {
+		// If getInstance errors, it means the host isn't ready.
+		// Return success with an empty list, as it's a query, not an operation.
+		logger.Warnf("[GO] ⚠️ Instance %d: GetConnectedPeers called but instance is not ready: %v\n", ni.instanceIndex, err)
+		return jsonSuccessResponse([]ExtendedPeerInfo{})
+	}
+
+	// Use a Write Lock for the entire critical section to avoid mixing RLock and Lock.
+	ni.peersMutex.RLock()
+	defer ni.peersMutex.RUnlock() // Ensure lock is released.
+
+	// Create a slice to hold the results directly from the map.
+	peersList := make([]ExtendedPeerInfo, 0, len(ni.friendlyPeers))
+	
+	for _, peerInfo := range ni.friendlyPeers {
+			peersList = append(peersList, peerInfo)
+		}
+
+	logger.Debugf("[GO] ℹ️ Instance %d: Reporting %d currently tracked and active peers.\n", ni.instanceIndex, len(peersList))
+
+	// Return the list of active peers as a JSON success response.
+	return jsonSuccessResponse(peersList) // Caller frees.
+}
+
+// GetRendezvousPeers returns a list of peers currently tracked as part of the world for a specific instance.
+// Note: This relies on the internal `rendezvousDiscoveredPeersInstances` map which is updated by pubsub
+// Parameters:
+//   - instanceIndexC (C.int): The index of the node instance.
+//
+// Returns:
+//   - *C.char: A JSON string containing a list of connected peers' information.
+//     Structure: `{"state":"Success", "message": [ExtendedPeerInfo, ...]}` or `{"state":"Error", "message":"..."}`.
+//     Each `ExtendedPeerInfo` object has `addr_info` (ID, Addrs), `connected_at`, `direction`, and `misc`.
+//   - IMPORTANT: The caller MUST free the returned C string using `FreeString`.
+//
+//export GetRendezvousPeers
+func GetRendezvousPeers(
+	instanceIndexC C.int,
+) *C.char {
+
+	ni, err := getInstance(int(instanceIndexC))
+	if err != nil {
+		// If instance isn't ready, we definitely don't have rendezvous peers.
+		return C.CString(`{"state":"Empty"}`)
+	}
+
+	ni.rendezvousMutex.RLock()
+	rendezvousState := ni.rendezvousState
+	ni.rendezvousMutex.RUnlock()
+
+	// If the state pointer is nil, it means we haven't received the first update yet.
+	if rendezvousState == nil {
+		return C.CString(`{"state":"Empty"}`)
+	}
+
+	// Extract the list of extendedPeerInfo to return it
+	peersList := make([]ExtendedPeerInfo, 0, len(rendezvousState.Peers))
+	for _, peerInfo := range rendezvousState.Peers {
+		peersList = append(peersList, peerInfo)
+	}
+
+	// This struct will be marshaled to JSON with exactly the fields you want.
+	responsePayload := struct {
+		Peers       []ExtendedPeerInfo `json:"peers"`
+		UpdateCount int64              `json:"update_count"`
+	}{
+		Peers:       peersList,
+		UpdateCount: rendezvousState.UpdateCount,
+	}
+
+	// The state exists, so return the whole struct.
+	logger.Debugf("[GO] ℹ️ Instance %d: Reporting %d rendezvous peers (UpdateCount: %d).\n", ni.instanceIndex, len(rendezvousState.Peers), rendezvousState.UpdateCount)
+	return jsonSuccessResponse(responsePayload) // Caller frees.
+}
+
+// GetNodeAddresses is the C-exported wrapper for goGetNodeAddresses.
+// It handles C-Go type conversions and JSON marshaling.
+//
+//export GetNodeAddresses
+func GetNodeAddresses(
+	instanceIndexC C.int,
+	peerIDC *C.char,
+) *C.char {
+
+	ni, err := getInstance(int(instanceIndexC))
+	if err != nil {
+		return jsonErrorResponse("Invalid instance", err)
+	}
+	peerIDStr := C.GoString(peerIDC) // Raw string from C
+	
+	var pidForInternalCall peer.ID // This will be peer.ID("") for local
+
+	if peerIDStr == "" || peerIDStr == ni.host.ID().String() {
+		// Convention: Empty peer.ID ("") passed to goGetNodeAddresses means "local node".
+		pidForInternalCall = "" // This is peer.ID("")
+	} else {
+		pidForInternalCall, err = peer.Decode(peerIDStr)
+		if err != nil {
+			errMsg := fmt.Sprintf("Instance %d: Failed to decode peer ID '%s'", ni.instanceIndex, peerIDStr)
+			return jsonErrorResponse(errMsg, err)
+		}
+	}
+
+	// Call the internal Go function with the resolved peer.ID or empty peer.ID for local
+	addresses, err := goGetNodeAddresses(ni, pidForInternalCall)
+	if err != nil {
+		return jsonErrorResponse(err.Error(), nil)
+	}
+
+	return jsonSuccessResponse(addresses)
+}
+
+// SendMessageToPeer sends a message either directly to a specific peer or broadcasts it via PubSub for a specific instance.
+// Parameters:
+//   - instanceIndexC (C.int): The index of the node instance.
+//   - channelC (*C.char): Use the unique channel as defined above in the Message struct.
+//   - dataC (*C.char): A pointer to the raw byte data of the message payload.
+//   - lengthC (C.int): The length of the data buffer pointed to by `data`.
+//
+// Returns:
+//   - *C.char: A JSON string with {"state": "Success/Error", "message": "..."}.
+//   - IMPORTANT: The caller MUST free this string using FreeString.
+//
+//export SendMessageToPeer
+func SendMessageToPeer(
+	instanceIndexC C.int,
+	channelC *C.char,
+	dataC *C.char,
+	lengthC C.int,
+) *C.char {
+
+	ni, err := getInstance(int(instanceIndexC))
+	if err != nil {
+		return jsonErrorResponse("Invalid instance", err)
+	}
+
+	// Convert C inputs
+	goChannel := C.GoString(channelC)
+	goData := C.GoBytes(unsafe.Pointer(dataC), C.int(lengthC))
+	
+	// --- Branch: Broadcast or Direct Send ---
+	if strings.Contains(goChannel, "::ps:") {
+		// --- Broadcast via specific PubSub Topic ---
+		instancePubsub := ni.pubsub
+		if instancePubsub == nil {
+			// PubSub not initialized, cannot broadcast
+			return jsonErrorResponse("PubSub not initialized, cannot broadcast", nil)
+		}
+
+		ni.pubsubMutex.Lock()
+		topic, exists := ni.topics[goChannel]
+		if !exists {
+			var err error
+			logger.Debugf("[GO]   - Instance %d: Joining PubSub topic '%s' for sending.\n", ni.instanceIndex, goChannel)
+			topic, err = instancePubsub.Join(goChannel) // ps is instancePubsub
+			if err != nil {
+				ni.pubsubMutex.Unlock()
+				// Failed to join PubSub topic
+				return jsonErrorResponse(fmt.Sprintf("Failed to join PubSub topic '%s'", goChannel), err)
+			}
+			ni.topics[goChannel] = topic
+			logger.Debugf("[GO] ✅ Instance %d: Joined PubSub topic: %s for publishing.\n", ni.instanceIndex, goChannel)
+		}
+		ni.pubsubMutex.Unlock()
+
+		// Directly publish the raw Protobuf payload.
+		if err := topic.Publish(ni.ctx, goData); err != nil {
+			// Failed to publish to topic
+			return jsonErrorResponse(fmt.Sprintf("Failed to publish to topic '%s'", goChannel), err)
+		}
+		logger.Infof("[GO] 🌍 Instance %d: Broadcast to topic '%s' (%d bytes)\n", ni.instanceIndex, goChannel, len(goData))
+		return jsonSuccessResponse(fmt.Sprintf("Message broadcast to topic %s", goChannel))
+
+	} else if strings.Contains(goChannel, "::dm:") {
+		// --- Direct Peer-to-Peer Message Sending (Persistent Stream Logic) ---
+		receiverChannelIDStr := strings.Split(goChannel, "::dm:")[1] // Extract the receiver's channel ID from the format "dm:<peerID>-<channelSpecifier>"
+		peerIDStr := strings.Split(receiverChannelIDStr, "-")[0]
+		pid, err := peer.Decode(peerIDStr)
+		if err != nil {
+			// Invalid peer ID format
+			return jsonErrorResponse("Invalid peer ID format in channel string", err)
+		}
+
+		if pid == ni.host.ID() {
+			// Attempt to send direct message to self
+			return jsonErrorResponse("Attempt to send direct message to self is invalid", nil)
+		}
+
+		ni.streamsMutex.Lock()
+		stream, streamExists := ni.persistentChatStreams[pid]
+		ni.streamsMutex.Unlock()
+
+		// If stream exists, try writing to it
+		if streamExists {
+			logger.Debugf("[GO]   ↳ Instance %d: Reusing stream %s to %s\n", ni.instanceIndex, stream.ID(), pid)
+			err = writeDirectMessageFrame(stream, goChannel, goData)
+			if err == nil {
+				logger.Infof("[GO] 📤 Instance %d: Sent to %s via Stream %s (Reused)\n", ni.instanceIndex, pid, stream.ID())
+				return jsonSuccessResponse(fmt.Sprintf("Direct message sent to %s (reused stream).", pid))
+			}
+			
+			// Write failed? Now we lock to remove the broken stream.
+			logger.Warnf("[GO] ⚠️ Instance %d: Write failed on Stream %s to %s: %v. Removing.\n", ni.instanceIndex, stream.ID(), pid, err)
+			ni.streamsMutex.Lock()
+			// Check if the stream in the map is still the broken one before deleting
+			if s, ok := ni.persistentChatStreams[pid]; ok && s == stream {
+				delete(ni.persistentChatStreams, pid)
+			}
+			ni.streamsMutex.Unlock()
+			_ = stream.Close() // Close the broken stream
+			return jsonErrorResponse(fmt.Sprintf("Failed to write to stream %s (closed).", pid), err)
+		} else {
+			// Stream does not exist, need to create a new one
+			logger.Debugf("[GO]   ↳ Instance %d: Creating NEW stream to %s...\n", ni.instanceIndex, pid)
+			streamCtx, cancel := context.WithTimeout(ni.ctx, 20*time.Second)
+			defer cancel()
+
+			newStream, err := ni.host.NewStream(
+				network.WithAllowLimitedConn(streamCtx, UnaiverseChatProtocol),
+				pid,
+				UnaiverseChatProtocol,
+			)
+
+			if err != nil {
+				return jsonErrorResponse(fmt.Sprintf("Failed to open new stream to %s.", pid), err)
+			}
+
+			// --- RACE CONDITION HANDLING ---
+			// Double-check if another goroutine created a stream while we were unlocked
+			ni.streamsMutex.Lock()
+			existingStream, existsNow := ni.persistentChatStreams[pid]
+			if existsNow {
+				logger.Warnf("[GO] ⚠️ Instance %d: Race detected. Using existing stream %s, closing our new %s.\n", ni.instanceIndex, existingStream.ID(), newStream.ID())
+				_ = newStream.Close() // Close the redundant stream we just created.
+				stream = existingStream
+			} else {
+				logger.Debugf("[GO] ✅ Instance %d: Opened and stored new persistent stream %s to %s\n", ni.instanceIndex, newStream.ID(), pid)
+				ni.persistentChatStreams[pid] = newStream
+				stream = newStream
+				go handleStream(ni, newStream)
+			}
+			ni.streamsMutex.Unlock()
+
+			// --- Write message to the determined stream ---
+			err = writeDirectMessageFrame(stream, goChannel, goData)
+			if err != nil {
+				logger.Errorf("[GO] ❌ Instance %d: Write failed on NEW stream %s to %s: %v.\n", ni.instanceIndex, stream.ID(), pid, err)
+				_ = stream.Close()
+				ni.streamsMutex.Lock()
+				if s, ok := ni.persistentChatStreams[pid]; ok && s == stream {
+					delete(ni.persistentChatStreams, pid)
+				}
+				ni.streamsMutex.Unlock()
+				return jsonErrorResponse(fmt.Sprintf("Failed to write to new stream to '%s' (needs reconnect).", pid), err)
+			}
+
+			logger.Infof("[GO] 📤 Instance %d: Sent to %s via Stream %s (New)\n", ni.instanceIndex, pid, stream.ID())
+			return jsonSuccessResponse(fmt.Sprintf("Direct message sent to %s (new stream).", pid))
+		}
+	} else {
+		// Invalid channel format
+		return jsonErrorResponse(fmt.Sprintf("Invalid channel format '%s'", goChannel), nil)
+	}
+}
+
+// SubscribeToTopic joins a PubSub topic and starts listening for messages for a specific instance.
+// Parameters:
+//   - instanceIndexC (C.int): The index of the node instance.
+//   - channelC (*C.char): The Channel associated to the topic to subscribe to.
+//
+// Returns:
+//   - *C.char: A JSON string indicating success or failure.
+//     Structure: `{"state":"Success", "message":"Subscribed to topic ..."}` or `{"state":"Error", "message":"..."}`.
+//   - IMPORTANT: The caller MUST free the returned C string using `FreeString`.
+//
+//export SubscribeToTopic
+func SubscribeToTopic(
+	instanceIndexC C.int,
+	channelC *C.char,
+) *C.char {
+
+	ni, err := getInstance(int(instanceIndexC))
+	if err != nil {
+		return jsonErrorResponse("Invalid instance", err)
+	}
+
+	// Convert C string input to Go string.
+	channel := C.GoString(channelC)
+	logger.Debugf("[GO] <sub> Instance %d: Attempting to subscribe to topic: %s\n", ni.instanceIndex, channel)
+	
+	// Get instance-specific state and mutex
+	instancePubsub := ni.pubsub
+	if ni.host == nil || instancePubsub == nil {
+		return jsonErrorResponse(
+			fmt.Sprintf("Instance %d: Host or PubSub not initialized", ni.instanceIndex), nil,
+		)
+	}
+
+	// Lock the mutex for safe access to the shared topics and subscriptions maps for this instance.
+	ni.pubsubMutex.Lock()
+	defer ni.pubsubMutex.Unlock() // Ensure mutex is unlocked when function returns.
+
+	// Check if already subscribed to this topic for this instance.
+	if _, exists := ni.subscriptions[channel]; exists {
+		logger.Debugf("[GO] <sub> Instance %d: Already subscribed to topic: %s\n", ni.instanceIndex, channel)
+		// Return success, indicating the desired state is already met.
+		return jsonSuccessResponse(
+			fmt.Sprintf("Instance %d: Already subscribed to topic %s", ni.instanceIndex, channel),
+		) // Caller frees.
+	}
+
+	// If the channel ends with ":rv", it indicates a rendezvous topic, so we remove other ones
+	// from the instanceTopics and instanceSubscriptions list, and we clean the rendezvousDiscoveredPeersInstances.
+	if strings.HasSuffix(channel, ":rv") {
+		logger.Debugf("  - Instance %d: Joining rendezvous topic '%s'. Cleaning up previous rendezvous state.\n", ni.instanceIndex, channel)
+		// Remove all existing rendezvous topics and subscriptions for this instance.
+		for existingChannel := range ni.topics {
+			if strings.HasSuffix(existingChannel, ":rv") {
+				logger.Debugf("  - Instance %d: Removing existing rendezvous topic '%s' from instance state.\n", ni.instanceIndex, existingChannel)
+
+				// Close the topic handle if it exists.
+				if topic, exists := ni.topics[existingChannel]; exists {
+					if err := topic.Close(); err != nil {
+						logger.Warnf("⚠️ Instance %d: Error closing topic handle for '%s': %v (proceeding with map cleanup)\n", ni.instanceIndex, existingChannel, err)
+					}
+					delete(ni.topics, existingChannel)
+				}
+
+				// Remove the subscription if it exists.
+				if sub, exists := ni.subscriptions[existingChannel]; exists {
+					sub.Cancel()                                   // Cancel the subscription
+					delete(ni.subscriptions, existingChannel) // Remove from map
+				}
+
+				// Also clean up rendezvous discovered peers for this instance.
+				logger.Debugf("  - Instance %d: Resetting rendezvous state for new topic '%s'.\n", ni.instanceIndex, channel)
+				ni.rendezvousMutex.Lock()
+				ni.rendezvousState = nil
+				ni.rendezvousMutex.Unlock()
+			}
+		}
+		logger.Debugf("  - Instance %d: Cleaned up previous rendezvous state.\n", ni.instanceIndex)
+	}
+
+	// --- Join the Topic ---
+	// Get a handle for the topic. `Join` creates the topic if it doesn't exist locally
+	// and returns a handle. It's safe to call Join multiple times; it's idempotent.
+	// We store the handle primarily for potential future publishing from this node.
+	topic, err := instancePubsub.Join(channel)
+	if err != nil {
+		errMsg := fmt.Sprintf("Instance %d: Failed to join topic '%s'", ni.instanceIndex, channel)
+		return jsonErrorResponse(errMsg, err) // Caller frees.
+	}
+	// Store the topic handle in the map for this instance.
+	ni.topics[channel] = topic
+	logger.Debugf("[GO]   - Instance %d: Obtained topic handle for: %s\n", ni.instanceIndex, channel)
+
+	// --- Subscribe to the Topic ---
+	// Create an actual subscription to receive messages from the topic.
+	sub, err := topic.Subscribe()
+	if err != nil {
+		// Close the newly created topic handle.
+		err := topic.Close()
+		if err != nil {
+			// Log error but proceed with cleanup.
+			logger.Warnf("[GO] ⚠️ Instance %d: Error closing topic handle for '%s': %v (proceeding with map cleanup)\n", ni.instanceIndex, channel, err)
+		}
+		// Remove the topic handle from our local map for this instance.
+		delete(ni.topics, channel)
+		errMsg := fmt.Sprintf("Instance %d: Failed to subscribe to topic '%s' after joining", ni.instanceIndex, channel)
+		return jsonErrorResponse(errMsg, err) // Caller frees.
+	}
+	// Store the subscription object in the map for this instance.
+	ni.subscriptions[channel] = sub
+	logger.Debugf("[GO]   - Instance %d: Created subscription object for: %s\n", ni.instanceIndex, channel)
+
+	// --- Start Listener Goroutine ---
+	// Launch a background goroutine that will continuously read messages
+	// from this new subscription and add them to the message buffer for this instance.
+	// Pass the instance index, subscription object, and topic name (for logging).
+	go readFromSubscription(ni, sub)
+
+	logger.Debugf("[GO] ✅ Instance %d: Subscribed successfully to topic: %s and started listener.\n", ni.instanceIndex, channel)
+	return jsonSuccessResponse(
+		fmt.Sprintf("Instance %d: Subscribed to topic %s", ni.instanceIndex, channel),
+	) // Caller frees.
+}
+
+// UnsubscribeFromTopic cancels an active PubSub subscription and cleans up related resources for a specific instance.
+// Parameters:
+//   - instanceIndexC (C.int): The index of the node instance.
+//   - channelC (*C.char): The Channel associated to the topic to unsubscribe from.
+//
+// Returns:
+//   - *C.char: A JSON string indicating success or failure.
+//     Structure: `{"state":"Success", "message":"Unsubscribed from topic ..."}` or `{"state":"Error", "message":"..."}`.
+//   - IMPORTANT: The caller MUST free the returned C string using `FreeString`.
+//
+//export UnsubscribeFromTopic
+func UnsubscribeFromTopic(
+	instanceIndexC C.int,
+	channelC *C.char,
+) *C.char {
+
+	ni, err := getInstance(int(instanceIndexC))
+	if err != nil {
+		// If instance is already gone, we can consider it "unsubscribed"
+		logger.Warnf("[GO] ⚠️ Instance %d: Unsubscribe called but instance is not ready: %v\n", ni.instanceIndex, err)
+		return jsonSuccessResponse(fmt.Sprintf("Instance %d: Not subscribed (instance not running)", ni.instanceIndex))
+	}
+
+	// Convert C string input to Go string.
+	channel := C.GoString(channelC)
+	logger.Debugf("[GO] </sub> Instance %d: Attempting to unsubscribe from topic: %s\n", ni.instanceIndex, channel)
+	
+	// Lock the mutex for write access to shared maps for this instance.
+	ni.pubsubMutex.Lock()
+	defer ni.pubsubMutex.Unlock()
+
+	// --- Cancel the Subscription ---
+	// Find the subscription object in the map for this instance.
+	sub, subExists := ni.subscriptions[channel]
+	if !subExists {
+		logger.Warnf("[GO] </sub> Instance %d: Not currently subscribed to topic: %s (or already unsubscribed)\n", ni.instanceIndex, channel)
+		// Also remove potential stale topic handle if subscription is gone.
+		delete(ni.topics, channel)
+		return jsonSuccessResponse(
+			fmt.Sprintf("Instance %d: Not currently subscribed to topic %s", ni.instanceIndex, channel),
+		) // Caller frees.
+	}
+
+	// Cancel the subscription. This signals the associated `readFromSubscription` goroutine
+	// (waiting on `sub.Next()`) to stop by causing `sub.Next()` to return an error (usually `ErrSubscriptionCancelled`).
+	// It also cleans up internal PubSub resources related to this subscription.
+	sub.Cancel()
+	// Remove the subscription entry from our local map for this instance.
+	delete(ni.subscriptions, channel)
+	logger.Debugf("[GO]   - Instance %d: Cancelled subscription object for topic: %s\n", ni.instanceIndex, channel)
+
+	// --- Close the Topic Handle ---
+	// Find the corresponding topic handle for this instance. It's good practice to close this as well,
+	// although PubSub might manage its lifecycle internally based on subscriptions.
+	// Explicit closing ensures resources related to the *handle* (like internal routing state) are released.
+	topic, topicExists := ni.topics[channel]
+	if topicExists {
+		logger.Debugf("[GO]   - Instance %d: Closing topic handle for: %s\n", ni.instanceIndex, channel)
+		// Close the topic handle.
+		err := topic.Close()
+		if err != nil {
+			// Log error but proceed with cleanup.
+			logger.Warnf("[GO] ⚠️ Instance %d: Error closing topic handle for '%s': %v (proceeding with map cleanup)\n", ni.instanceIndex, channel, err)
+		}
+		// Remove the topic handle from our local map for this instance.
+		delete(ni.topics, channel)
+		logger.Debugf("[GO]   - Instance %d: Removed topic handle from local map for topic: %s\n", ni.instanceIndex, channel)
+	} else {
+		logger.Debugf("[GO]   - Instance %d: No topic handle found in local map for '%s' to close (already removed or possibly never stored?).\n", ni.instanceIndex, channel)
+		// Ensure removal from map even if handle wasn't found (e.g., inconsistent state).
+		delete(ni.topics, channel)
+	}
+
+	// If the channel ends with ":rv", it indicates a rendezvous topic, so we have closed the topic and the sub
+	// but we also need to clean the rendezvousDiscoveredPeersInstances.
+	if strings.HasSuffix(channel, ":rv") {
+		logger.Debugf("  - Instance %d: Unsubscribing from rendezvous topic. Clearing state.\n", ni.instanceIndex)
+		ni.rendezvousMutex.Lock()
+		ni.rendezvousState = nil
+		ni.rendezvousMutex.Unlock()
+	}
+	logger.Debugf("[GO]   - Instance %d: Cleaned up previous rendezvous state.\n", ni.instanceIndex)
+
+	logger.Infof("[GO] ✅ Instance %d: Unsubscribed successfully from topic: %s\n", ni.instanceIndex, channel)
+	return jsonSuccessResponse(
+		fmt.Sprintf("Instance %d: Unsubscribed from topic %s", ni.instanceIndex, channel),
+	) // Caller frees.
+}
+
+// MessageQueueLength returns the total number of messages waiting across all channel queues for a specific instance.
+// Parameters:
+//   - instanceIndexC (C.int): The index of the node instance.
+//
+// Returns:
+//   - C.int: The total number of messages. Returns -1 if instance index is invalid.
+//
+//export MessageQueueLength
+func MessageQueueLength(
+	instanceIndexC C.int,
+) C.int {
+
+	ni, err := getInstance(int(instanceIndexC))
+	if err != nil {
+		logger.Errorf("[GO] ❌ MessageQueueLength: %v\n", err)
+		return -1 // Return -1 if instance isn't valid
+	}
+
+	// Get the message store for this instance
+	store := ni.messageStore
+	if store == nil {
+		logger.Errorf("[GO] ❌ Instance %d: Message store not initialized.\n", ni.instanceIndex)
+		return 0 // Return 0 if store is nil (effectively empty)
+	}
+
+	store.mu.Lock()
+	defer store.mu.Unlock()
+
+	totalLength := 0
+	// TODO: this makes sense but not for the check we are doing from python, think about it
+	for _, messageList := range store.messagesByChannel {
+		totalLength += messageList.Len()
+	}
+
+	return C.int(totalLength)
+}
+
+// PopMessages retrieves the oldest message from each channel's queue for a specific instance.
+// This function always pops one message per channel that has messages.
+// Parameters:
+//   - instanceIndexC (C.int): The index of the node instance.
+//
+// Returns:
+//   - *C.char: A JSON string representing a list of the popped messages.
+//     Returns `{"state":"Empty"}` if no messages were available in any queue.
+//     Returns `{"state":"Error", "message":"..."}` on failure.
+//   - IMPORTANT: The caller MUST free the returned C string using `FreeString`.
+//
+//export PopMessages
+func PopMessages(
+	instanceIndexC C.int,
+) *C.char {
+
+	ni, err := getInstance(int(instanceIndexC))
+	if err != nil {
+		return jsonErrorResponse("Invalid instance", err)
+	}
+
+	// Get the message store for this instance
+	store := ni.messageStore
+	if store == nil {
+		logger.Errorf("[GO] ❌ Instance %d: PopMessages: Message store not initialized.\n", ni.instanceIndex)
+		return jsonErrorResponse(fmt.Sprintf("Instance %d: Message store not initialized", ni.instanceIndex), nil)
+	}
+
+	store.mu.Lock() // Lock for the entire operation
+	defer store.mu.Unlock()
+
+	if len(store.messagesByChannel) == 0 {
+		return C.CString(`{"state":"Empty"}`)
+	}
+
+	// Create a slice to hold the popped messages. Capacity is the number of channels.
+	var poppedMessages []*QueuedMessage
+	for channel, messageList := range store.messagesByChannel {
+		if messageList.Len() > 0 {
+			element := messageList.Front()
+			msg := element.Value.(*QueuedMessage)
+			poppedMessages = append(poppedMessages, msg)
+			messageList.Remove(element)
+		}
+		// if the queue is now empty, we can delete it from the map to save space
+		if messageList.Len() == 0 {
+			delete(store.messagesByChannel, channel)
+		}
+	}
+
+	// After iterating, check if we actually popped anything
+	if len(poppedMessages) == 0 {
+		return C.CString(`{"state":"Empty"}`)
+	}
+
+	// Marshal the slice of popped messages into a JSON array.
+	// We create a temporary structure for JSON marshalling to include the base64-encoded data.
+	payloads := make([]map[string]interface{}, len(poppedMessages))
+	for i, msg := range poppedMessages {
+		payloads[i] = map[string]interface{}{
+			"from": msg.From,
+			"data": base64.StdEncoding.EncodeToString(msg.Data),
+		}
+	}
+
+	jsonBytes, err := json.Marshal(payloads)
+	if err != nil {
+		logger.Errorf("[GO] ❌ Instance %d: PopMessages: Failed to marshal messages to JSON: %v\n", ni.instanceIndex, err)
+		// Messages have already been popped from the queue at this point.
+		// Returning an error is the best we can do.
+		return jsonErrorResponse(
+			fmt.Sprintf("Instance %d: Failed to marshal popped messages", ni.instanceIndex), err,
+		)
+	}
+
+	return C.CString(string(jsonBytes))
+}
+
+// CloseNode gracefully shuts down the libp2p host, cancels subscriptions, closes connections,
+// and cleans up all associated resources.
+// Parameters:
+//   - instanceIndexC (C.int): The index of the node instance. If -1, closes all initialized instances.
+//
+// Returns:
+//   - *C.char: A JSON string indicating the result of the closure attempt.
+//     Structure: `{"state":"Success", "message":"Node closed successfully"}` or `{"state":"Error", "message":"Error closing host: ..."}`.
+//     If closing all, the message will summarize the results.
+//   - IMPORTANT: The caller MUST free the returned C string using `FreeString`.
+//
+//export CloseNode
+func CloseNode(
+	instanceIndexC C.int,
+) *C.char {
+	
+	instanceIndex := int(instanceIndexC)
+
+	if instanceIndex == -1 {
+		logger.Debugf("[GO] 🛑 Closing all initialized instances of this node...")
+		successCount := 0
+		errorCount := 0
+		var errorMessages []string
+
+		// acquire the global lock
+		globalInstanceMutex.Lock()
+		defer globalInstanceMutex.Unlock()
+
+		for i, ni := range allInstances {
+			if ni != nil {
+				logger.Debugf("[GO] 🛑 Attempting to close instance %d...\n", i)
+				
+				err := ni.Close() // Call the new method
+				allInstances[i] = nil // Remove from slice
+				
+				if err != nil {
+					errorCount++
+					errorMessages = append(errorMessages, fmt.Sprintf("Instance %d: %v", i, err))
+					logger.Errorf("[GO] ❌ Instance %d: Close failed: %v\n", i, err)
+				} else {
+					successCount++
+					logger.Debugf("[GO] ✅ Instance %d: Closed successfully.\n", i)
+				}
+			}
+		}
+
+		summaryMsg := fmt.Sprintf("Closed %d nodes successfully, %d failed.", successCount, errorCount)
+		if errorCount > 0 {
+			logger.Errorf("[GO] ❌ Errors encountered during batch close:\n")
+			for _, msg := range errorMessages {
+				logger.Errorf(msg)
+			}
+			return jsonErrorResponse(summaryMsg, fmt.Errorf("details: %v", errorMessages))
+		}
+
+		logger.Infof("[GO] 🛑 All initialized nodes closed.")
+		return jsonSuccessResponse(summaryMsg)
+
+	} else {
+		if instanceIndex < 0 || instanceIndex >= maxInstances {
+			err := fmt.Errorf("invalid instance index: %d. Must be between 0 and %d", instanceIndex, maxInstances-1)
+			return jsonErrorResponse("Invalid instance index for single close", err) // Caller frees.
+		}
+
+		globalInstanceMutex.Lock()
+		defer globalInstanceMutex.Unlock()
+
+		instance := allInstances[instanceIndex]
+		if instance == nil {
+			logger.Debugf("[GO] ℹ️ Instance %d: Node was already closed.\n", instanceIndex)
+			return jsonSuccessResponse(fmt.Sprintf("Instance %d: Node was already closed", instanceIndex))
+		}
+
+		err := instance.Close()
+		allInstances[instanceIndex] = nil
+
+		if err != nil {
+			return jsonErrorResponse(fmt.Sprintf("Instance %d: Error closing host", instanceIndex), err)
+		}
+
+		logger.Infof("[GO] 🛑 Instance %d: Node closed successfully.\n", instanceIndex)
+		return jsonSuccessResponse(fmt.Sprintf("Instance %d: Node closed successfully", instanceIndex))
+	}
+}
+
+// FreeString is called from the C/Python side to release the memory allocated by Go
+// when returning a `*C.char` (via `C.CString`).
+// Parameters:
+//   - s (*C.char): The pointer to the C string previously returned by an exported Go function.
+//
+//export FreeString
+func FreeString(
+	s *C.char,
+) {
+
+	// Check for NULL pointer before attempting to free.
+	if s != nil {
+		C.free(unsafe.Pointer(s)) // Use C.free via unsafe.Pointer to release the memory.
+	}
+}
+
+// FreeInt is provided for completeness but is generally **NOT** needed if Go functions
+// only return `C.int` (by value). It would only be necessary if a Go function manually
+// allocated memory for a C integer (`*C.int`) and returned the pointer, which is uncommon.
+// Parameters:
+//   - i (*C.int): The pointer to the C integer previously allocated and returned by Go.
+//
+//export FreeInt
+func FreeInt(
+	i *C.int,
+) {
+
+	// Check for NULL pointer.
+	if i != nil {
+		logger.Warnf("[GO] ⚠️ FreeInt called - Ensure a *C.int pointer was actually allocated and returned from Go (this is unusual).")
+		C.free(unsafe.Pointer(i)) // Free the memory if it was indeed allocated.
+	}
+}
+
+// main is the entry point for a Go executable.
+func main() {
+	// This message will typically only be seen if you run `go run lib.go`
+	// or build and run as a standard executable, NOT when used as a shared library.
+	logger.Debugf("[GO] libp2p Go library main function (not executed in c-shared library mode)")
+}