Feature: Web remote desktop bidirectional text clipboard sync

Add keyboard-shortcut (Ctrl+C / Ctrl+X / Ctrl+V) text clipboard transfer
between the browser and the remote desktop, reusing the clipboard protocol
that already exists in the native viewer (COMMAND_SCREEN_GET_CLIPBOARD=24,
COMMAND_SCREEN_SET_CLIPBOARD=25, TOKEN_CLIPBOARD_TEXT=118) so the client
needs no change.

Local -> remote: pressing Ctrl+V reads the browser clipboard, pushes it to
the remote clipboard (UTF-8 -> GBK, matching the client's CF_TEXT/ANSI
handling) and then synthesizes a remote Ctrl+V so the pasted text is the
newly delivered one.

Remote -> local: Ctrl+C or Ctrl+X is forwarded to the remote as usual and,
after a short delay, the remote clipboard is pulled back and written to the
browser clipboard for a later local Ctrl+V.

Both the MFC and Go servers get the same three clipboard signaling paths
(clipboard_set, clipboard_get, clipboard_text) so the shared single-page
frontend behaves identically on either backend.

Co-Authored-By: deepseek-v4-pro

server/2015Remote/ScreenSpyDlg.cpp

@@ -1126,6 +1126,10 @@ VOID CScreenSpyDlg::OnReceiveComplete()
     case TOKEN_CLIPBOARD_TEXT: {
         Buffer str = m_ContextObject->InDeCompressedBuffer.GetMyBuffer(1);
         UpdateServerClipboard(str.c_str(), str.length());
+        // 反向回传：远程剪贴板文本（UTF-8）→ Web 客户端
+        if (m_bIsWebSession && WebService().IsRunning()) {
+            WebService().NotifyClipboardText(m_ClientID, std::string(str.c_str(), str.length()));
+        }
         break;
     }
     case TOKEN_BITMAPINFO: {

server/2015Remote/WebService.cpp

@@ -393,6 +393,10 @@ void CWebService::ServerThread(int port) {
                 HandleMouse(ws_ptr, msg);
             } else if (cmd == "key") {
                 HandleKey(ws_ptr, msg);
+            } else if (cmd == "clipboard_set") {
+                HandleClipboardSet(ws_ptr, msg);
+            } else if (cmd == "clipboard_get") {
+                HandleClipboardGet(ws_ptr, token);
             } else if (cmd == "rdp_reset") {
                 HandleRdpReset(ws_ptr, token);
             } else if (cmd == "audio_toggle") {
@@ -929,6 +933,93 @@ void CWebService::HandleKey(void* ws_ptr, const std::string& msg) {
     ctx->Send2Client(szData, length);
 }
 
+void CWebService::HandleClipboardSet(void* ws_ptr, const std::string& msg) {
+    // Parse JSON
+    Json::Value root;
+    Json::Reader reader;
+    if (!reader.parse(msg, root)) {
+        return;
+    }
+
+    std::string token = root.get("token", "").asString();
+    std::string text = root.get("text", "").asString();
+
+    if (text.empty()) return;
+
+    // Validate token
+    std::string username, role;
+    if (!ValidateToken(token, username, role)) {
+        return;
+    }
+
+    // Get device being watched
+    uint64_t device_id = 0;
+    {
+        std::lock_guard<std::mutex> lock(m_ClientsMutex);
+        auto it = m_Clients.find(ws_ptr);
+        if (it != m_Clients.end()) {
+            device_id = it->second.watch_device_id;
+            it->second.last_activity = (uint64_t)time(nullptr);
+        }
+    }
+
+    if (device_id == 0) return;
+
+    // Get screen context
+    CONTEXT_OBJECT* ctx = GetScreenContext(device_id);
+    if (!ctx) {
+        return;
+    }
+
+    // UTF-8 → UTF-16 → GBK(936)。客户端 UpdateClientClipboard 以 CF_TEXT(ANSI/GBK) 落剪贴板，
+    // 故显式用 936 而非 CP_ACP，避免服务端 locale 非中文时出错。
+    int wlen = MultiByteToWideChar(CP_UTF8, 0, text.c_str(), (int)text.size(), nullptr, 0);
+    if (wlen <= 0) return;
+    std::wstring wide(wlen, L'\0');
+    MultiByteToWideChar(CP_UTF8, 0, text.c_str(), (int)text.size(), &wide[0], wlen);
+
+    int glen = WideCharToMultiByte(936, 0, wide.c_str(), wlen, nullptr, 0, nullptr, nullptr);
+    if (glen <= 0) return;
+
+    // 命令字节 + GBK 文本
+    std::vector<BYTE> szData(1 + glen);
+    szData[0] = COMMAND_SCREEN_SET_CLIPBOARD;
+    WideCharToMultiByte(936, 0, wide.c_str(), wlen, (char*)szData.data() + 1, glen, nullptr, nullptr);
+
+    ctx->Send2Client(szData.data(), (ULONG)szData.size());
+}
+
+void CWebService::HandleClipboardGet(void* ws_ptr, const std::string& token) {
+    // Validate token
+    std::string username, role;
+    if (!ValidateToken(token, username, role)) {
+        return;
+    }
+
+    // Get device being watched
+    uint64_t device_id = 0;
+    {
+        std::lock_guard<std::mutex> lock(m_ClientsMutex);
+        auto it = m_Clients.find(ws_ptr);
+        if (it != m_Clients.end()) {
+            device_id = it->second.watch_device_id;
+            it->second.last_activity = (uint64_t)time(nullptr);
+        }
+    }
+
+    if (device_id == 0) return;
+
+    // Get screen context
+    CONTEXT_OBJECT* ctx = GetScreenContext(device_id);
+    if (!ctx) {
+        return;
+    }
+
+    // 单字节命令，客户端以 SendClientClipboard → TOKEN_CLIPBOARD_TEXT 回包
+    BYTE bToken = COMMAND_SCREEN_GET_CLIPBOARD;
+    ctx->Send2Client(&bToken, sizeof(BYTE));
+}
+
 void CWebService::HandleRdpReset(void* ws_ptr, const std::string& token) {
     std::string username, role;
     if (!ValidateToken(token, username, role)) {
@@ -1711,6 +1802,26 @@ void CWebService::BroadcastCursor(uint64_t device_id, uint8_t cursor_index) {
     }
 }
 
+void CWebService::NotifyClipboardText(uint64_t device_id, const std::string& text) {
+    if (m_bStopping) return;
+
+    Json::Value res;
+    res["cmd"] = "clipboard_text";
+    res["id"] = device_id;
+    res["text"] = text;
+
+    Json::StreamWriterBuilder builder;
+    builder["indentation"] = "";
+    std::string json = Json::writeString(builder, res);
+
+    std::lock_guard<std::mutex> lock(m_ClientsMutex);
+    for (auto& [ws_ptr, client] : m_Clients) {
+        if (client.watch_device_id == device_id) {
+            SendText(ws_ptr, json);
+        }
+    }
+}
+
 bool CWebService::StartRemoteDesktop(uint64_t device_id) {
     if (!m_pParentDlg) return false;
 

server/2015Remote/WebService.h

@@ -160,6 +160,9 @@ public:
     // Cursor change notification (called from ScreenSpyDlg)
     void BroadcastCursor(uint64_t device_id, uint8_t cursor_index);
 
+    // 远程剪贴板文本回传通知（called from ScreenSpyDlg，payload 为 UTF-8）
+    void NotifyClipboardText(uint64_t device_id, const std::string& text);
+
     // Get count of web clients watching a device
     int GetWebClientCount(uint64_t device_id);
 
@@ -187,6 +190,8 @@ private:
     void HandlePing(void* ws_ptr, const std::string& token);
     void HandleMouse(void* ws_ptr, const std::string& msg);
     void HandleKey(void* ws_ptr, const std::string& msg);
+    void HandleClipboardSet(void* ws_ptr, const std::string& msg);
+    void HandleClipboardGet(void* ws_ptr, const std::string& token);
     void HandleRdpReset(void* ws_ptr, const std::string& token);
     void HandleAudioToggle(void* ws_ptr, const std::string& token);
 

server/go/cmd/main.go

@@ -135,6 +135,8 @@ func (h *MyHandler) OnReceive(ctx *connection.Context, data []byte) {
 		// arrives shortly after via TOKEN_NEXTSCREEN.
 	case protocol.TokenNextScreen:
 		h.handleScreenFrame(ctx, data, false)
+	case protocol.TokenClipboardText:
+		h.handleClipboardText(ctx, data)
 	case protocol.TokenKeyframe:
 		// Sent by the client only when frameID % m_GOP == 0; the client's
 		// DEFAULT_GOP is 0x7FFFFFFF (effectively infinite), so this token
@@ -427,6 +429,18 @@ func (h *MyHandler) handleScreenFrame(ctx *connection.Context, data []byte, alwa
 	h.hub.PublishScreenFrame(deviceID, packet, isKey)
 }
 
+// handleClipboardText handles TOKEN_CLIPBOARD_TEXT (118): the device replies to
+// a COMMAND_SCREEN_GET_CLIPBOARD request with its clipboard text as
+// [118][UTF-8 text][NUL]. data[0] is the token; the payload is data[1:].
+func (h *MyHandler) handleClipboardText(ctx *connection.Context, data []byte) {
+	deviceID := h.hub.ScreenDeviceID(ctx)
+	if deviceID == "" {
+		return // not a bound screen sub-conn — drop
+	}
+	text := strings.TrimRight(string(data[1:]), "\x00")
+	h.hub.PublishClipboardText(deviceID, text)
+}
+
 // handleLogin handles client login (TOKEN_LOGIN = 102)
 func (h *MyHandler) handleLogin(ctx *connection.Context, data []byte) {
 	info, err := protocol.ParseLoginInfo(data)

server/go/hub/hub.go

@@ -187,6 +187,9 @@ type EventHandler interface {
 	// Duplicates (same index as the previous frame) are filtered out by the
 	// hub before reaching subscribers.
 	OnCursorChange(deviceID string, index byte)
+	// OnClipboardText delivers the device's clipboard text (already UTF-8),
+	// captured in response to a browser "clipboard_get" request.
+	OnClipboardText(deviceID string, text string)
 	// OnTerminalReady fires once the device's shell sub-conn is bound and
 	// the server has sent COMMAND_NEXT to start its output read loop.
 	// isPTY=true means PTY mode (Linux/macOS or ConPTY); false means the
@@ -606,6 +609,14 @@ func (h *Hub) PublishScreenFrame(deviceID string, packet []byte, isKeyframe bool
 	}
 }
 
+// PublishClipboardText relays a device's clipboard text (UTF-8) to all
+// subscribers, e.g. in response to a browser "clipboard_get" request.
+func (h *Hub) PublishClipboardText(deviceID string, text string) {
+	for _, s := range h.snapshotSubscribers() {
+		s.OnClipboardText(deviceID, text)
+	}
+}
+
 // UpdateLive applies a heartbeat-derived RTT and active-window title to the
 // device's live fields, then notifies subscribers. No-op if the device is
 // not registered (e.g. heartbeat arriving for a connection that never sent

server/go/hub/hub_test.go

@@ -84,6 +84,8 @@ func (c *captureHandler) OnResolutionChange(_ string, _, _ int) {}
 
 func (c *captureHandler) OnCursorChange(_ string, _ byte) {}
 
+func (c *captureHandler) OnClipboardText(_ string, _ string) {}
+
 func (c *captureHandler) OnTerminalReady(_ string, _ bool) {}
 
 func (c *captureHandler) OnTerminalData(_ string, _ []byte) {}

server/go/protocol/commands.go

@@ -103,15 +103,17 @@ func DecodeClientString(data []byte, capability, clientType string) string {
 // Command tokens - matching the C++ definitions (common/commands.h).
 const (
 	// Server -> Client commands
-	CommandActived       byte = 0   // COMMAND_ACTIVED
+	CommandActived            byte = 0   // COMMAND_ACTIVED
-	CommandScreenSpy     byte = 16  // COMMAND_SCREEN_SPY - start screen capture
+	CommandScreenSpy          byte = 16  // COMMAND_SCREEN_SPY - start screen capture
-	CommandScreenControl byte = 20  // COMMAND_SCREEN_CONTROL - mouse/keyboard input (MSG64 batches)
+	CommandScreenControl      byte = 20  // COMMAND_SCREEN_CONTROL - mouse/keyboard input (MSG64 batches)
-	CommandNext          byte = 30  // COMMAND_NEXT - "control-side dialog is open, you may stream"
+	CommandScreenGetClipboard byte = 24  // COMMAND_SCREEN_GET_CLIPBOARD - request remote clipboard (single byte)
-	CommandShell         byte = 40  // COMMAND_SHELL - ask device to open a shell sub-connection
+	CommandScreenSetClipboard byte = 25  // COMMAND_SCREEN_SET_CLIPBOARD - set remote clipboard [25][GBK text]
-	CommandTerminalRsize byte = 81  // CMD_TERMINAL_RESIZE - [cmd:1][cols:2 LE][rows:2 LE]
+	CommandNext               byte = 30  // COMMAND_NEXT - "control-side dialog is open, you may stream"
-	CmdRestoreConsole    byte = 82  // CMD_RESTORE_CONSOLE - RDP session "归位": switch back to the console session and restart capture
+	CommandShell              byte = 40  // COMMAND_SHELL - ask device to open a shell sub-connection
-	CommandBye           byte = 204 // COMMAND_BYE - disconnect
+	CommandTerminalRsize      byte = 81  // CMD_TERMINAL_RESIZE - [cmd:1][cols:2 LE][rows:2 LE]
-	CommandHeartbeat     byte = 216 // CMD_HEARTBEAT_ACK
+	CmdRestoreConsole         byte = 82  // CMD_RESTORE_CONSOLE - RDP session "归位": switch back to the console session and restart capture
+	CommandBye                byte = 204 // COMMAND_BYE - disconnect
+	CommandHeartbeat          byte = 216 // CMD_HEARTBEAT_ACK
 
 	// Client -> Server tokens
 	TokenAuth          byte = 100 // TOKEN_AUTH - authorization required
@@ -120,6 +122,7 @@ const (
 	TokenBitmapInfo    byte = 115 // TOKEN_BITMAPINFO - screen sub-connection header
 	TokenFirstScreen   byte = 116 // TOKEN_FIRSTSCREEN - raw BGRA baseline frame (NOT H264)
 	TokenNextScreen    byte = 117 // TOKEN_NEXTSCREEN - non-keyframe H264 (P-frame)
+	TokenClipboardText byte = 118 // TOKEN_CLIPBOARD_TEXT - remote clipboard text [118][UTF-8][NUL]
 	TokenShellStart    byte = 128 // TOKEN_SHELL_START - legacy cmd-pipe shell sub-conn open
 	TokenKeyframe      byte = 134 // TOKEN_KEYFRAME - H264 IDR (sent on GOP boundary)
 	TokenTerminalStart byte = 232 // TOKEN_TERMINAL_START - modern PTY shell sub-conn open

server/go/web/ws.go

@@ -114,9 +114,9 @@ type wsHub struct {
 
 func newWSHub(auth *wsauth.Authenticator, devices *hub.Hub, log *logger.Logger) *wsHub {
 	h := &wsHub{
-		auth:    auth,
+		auth:              auth,
-		devices: devices,
+		devices:           devices,
-		log:     log,
+		log:               log,
 		clients:           make(map[*wsClient]struct{}),
 		screenCloseTimers: make(map[string]*time.Timer),
 	}
@@ -300,6 +300,23 @@ func (h *wsHub) OnResolutionChange(deviceID string, width, height int) {
 	}
 }
 
+// OnClipboardText relays a device's clipboard text back to every browser
+// watching it (in response to a "clipboard_get" request).
+func (h *wsHub) OnClipboardText(deviceID string, text string) {
+	msg := mustJSON(map[string]any{
+		"cmd":  "clipboard_text",
+		"id":   deviceID,
+		"text": text,
+	})
+	h.mu.RLock()
+	defer h.mu.RUnlock()
+	for c := range h.clients {
+		if c.watching == deviceID && c.token != "" {
+			c.queue(msg)
+		}
+	}
+}
+
 // OnScreenFrame ships a screen packet to every browser currently watching
 // this device. We hold the read lock for the whole iteration, but each
 // queueBinary is non-blocking (drops on backpressure) so a slow viewer

server/go/web/ws_handlers.go

@@ -8,6 +8,7 @@ import (
 	"github.com/yuanyuanxiang/SimpleRemoter/server/go/hub"
 	"github.com/yuanyuanxiang/SimpleRemoter/server/go/protocol"
 	"github.com/yuanyuanxiang/SimpleRemoter/server/go/wsauth"
+	"golang.org/x/text/encoding/simplifiedchinese"
 )
 
 // rotateChallenge replaces the client's nonce with a fresh one and pushes a
@@ -64,6 +65,10 @@ func (h *wsHub) dispatch(c *wsClient, cmd string, raw []byte) {
 		h.handleMouse(c, raw)
 	case "key":
 		h.handleKey(c, raw)
+	case "clipboard_set":
+		h.handleClipboardSet(c, raw)
+	case "clipboard_get":
+		h.handleClipboardGet(c, raw)
 	case "term_open":
 		h.handleTermOpen(c, raw)
 	case "term_input":
@@ -559,6 +564,59 @@ func (h *wsHub) handleKey(c *wsClient, raw []byte) {
 	_ = h.devices.SendToScreen(deviceID, pkt)
 }
 
+// handleClipboardSet relays the browser's local clipboard text into the remote
+// device's clipboard via COMMAND_SCREEN_SET_CLIPBOARD (25). The client's
+// UpdateClientClipboard stores the payload as CF_TEXT (ANSI/GBK), so we encode
+// the UTF-8 input to GBK here — matching the C++ WebService (explicit code
+// page 936) and the existing MCP remote_clipboard convention. Non-GBK glyphs
+// (emoji, some accented Latin) collapse to '?' — an accepted limitation.
+func (h *wsHub) handleClipboardSet(c *wsClient, raw []byte) {
+	if !h.requireAuth(c, raw, "clipboard_set_result") {
+		return
+	}
+	var in struct {
+		Text string `json:"text"`
+	}
+	if err := json.Unmarshal(raw, &in); err != nil {
+		return
+	}
+	if in.Text == "" {
+		return
+	}
+	deviceID := c.watching
+	if deviceID == "" {
+		return
+	}
+	gbk, err := simplifiedchinese.GBK.NewEncoder().Bytes([]byte(in.Text))
+	if err != nil {
+		h.log.Warn("clipboard_set: GBK encode failed: %v", err)
+		return
+	}
+	payload := append([]byte{protocol.CommandScreenSetClipboard}, gbk...)
+	if err := h.devices.SendToScreen(deviceID, payload); err != nil {
+		h.log.Warn("clipboard_set: device=%s: %v", deviceID, err)
+		return
+	}
+	h.log.Info("clipboard_set sent: device=%s len=%d", deviceID, len(gbk))
+}
+
+// handleClipboardGet asks the remote device for its clipboard text via
+// COMMAND_SCREEN_GET_CLIPBOARD (24, a single byte). The device replies with
+// TOKEN_CLIPBOARD_TEXT, which the hub relays back to the browser.
+func (h *wsHub) handleClipboardGet(c *wsClient, raw []byte) {
+	if !h.requireAuth(c, raw, "clipboard_get_result") {
+		return
+	}
+	deviceID := c.watching
+	if deviceID == "" {
+		return
+	}
+	if err := h.devices.SendToScreen(deviceID, []byte{protocol.CommandScreenGetClipboard}); err != nil {
+		h.log.Warn("clipboard_get: device=%s: %v", deviceID, err)
+		return
+	}
+}
+
 // handleCreateUser provisions a new web account. Admin-only. Mirrors the
 // C++ CWebService::HandleCreateUser semantics (WebService.cpp:1009): the
 // password is hashed with a fresh salt, the user table is persisted, and

server/web/index.html

@@ -1365,6 +1365,7 @@
     <script>
         let ws = null, token = null, decoder = null, devices = [], currentDevice = null;
         let frameCount = 0, lastFrameTime = 0, fps = 0, pingInterval = null;
+        let skipNextCtrlVKeyup = false;   // 剪贴板桥：抑制 pasteToRemote 合成的 Ctrl+V 物理 keyup
         // FPS 计数：原始风格——decoder.onOutput 里 ++，每经过 1 秒采样一次。
         // 简单直接，与本次会话改动前一致。
         // 网络流量统计：handleBinaryFrame 累加，每 1 秒钟 renderStats 读出
@@ -1863,6 +1864,12 @@
                         applyRemoteCursor(msg.index);
                     }
                     break;
+                case 'clipboard_text':
+                    // 远程剪贴板文本回传：写回浏览器（即本机）剪贴板
+                    if (msg.text) {
+                        try { navigator.clipboard.writeText(msg.text); } catch (e) {}
+                    }
+                    break;
                 case 'device_offline':
                     // Only handle if this is the device we're currently viewing
                     if (!token) break;
@@ -3714,6 +3721,28 @@
             }
         }
 
+        // 本地剪贴板 → 远程：读浏览器剪贴板，先 clipboard_set 到远程，再合成一次完整 Ctrl+V。
+        // 合成按键放在 readText 的 then 里，保证 clipboard_set 先于 Ctrl+V 到达（同一条 WS 有序）。
+        function pasteToRemote() {
+            const press = () => { sendKey(86, true, false); sendKey(86, false, false); };
+            navigator.clipboard.readText().then(function(text) {
+                if (text && ws && ws.readyState === WebSocket.OPEN && token) {
+                    ws.send(JSON.stringify({ cmd: 'clipboard_set', token, text }));
+                }
+                press();
+            }).catch(function() { press(); });  // 读剪贴板被拒/为空：仍执行一次普通远程粘贴
+        }
+
+        // 远程剪贴板 → 本地：等远程完成拷贝（启发式延时）后请求远程剪贴板，回包写回浏览器剪贴板。
+        // 不保证 100% 命中，失败用户二次 Ctrl+C 即可。
+        function scheduleCopyFromRemote() {
+            setTimeout(function() {
+                if (ws && ws.readyState === WebSocket.OPEN && token) {
+                    ws.send(JSON.stringify({ cmd: 'clipboard_get', token }));
+                }
+            }, 250);
+        }
+
         function showTouchIndicator(x, y) {
             touchIndicator.style.left = x + 'px';
             touchIndicator.style.top = y + 'px';
@@ -4198,14 +4227,38 @@
             // controlling the remote — see the F11 handler above. Skip here
             // to let that one handle it.
             if (e.key === 'F11' && !controlEnabled) return;
+
+            // 文本剪贴板桥（secure context 下 navigator.clipboard 可用）。
+            // Ctrl/Cmd+V：把本地剪贴板文本先灌到远程剪贴板，再合成一次 Ctrl+V 粘贴。
+            // Ctrl/Cmd+C / Ctrl/Cmd+X：照常转发给远程（远程完成拷贝/剪切），随后延迟拉取远程剪贴板回本地。
+            if (controlEnabled && (e.ctrlKey || e.metaKey) && !e.shiftKey && !e.altKey) {
+                if (e.keyCode === 86) {            // Ctrl/Cmd+V
+                    e.preventDefault();
+                    skipNextCtrlVKeyup = true;
+                    pasteToRemote();
+                    return;
+                }
+                // e.keyCode === 67(C)/88(X) 时落到下面的 sendKey 原样转发，转发后再延迟拉取
+            }
+
             e.preventDefault();
             sendKey(e.keyCode, true, e.altKey);
+
+            if (controlEnabled && (e.ctrlKey || e.metaKey) && !e.shiftKey && !e.altKey && (e.keyCode === 67 || e.keyCode === 88)) {
+                scheduleCopyFromRemote();
+            }
         });
 
         document.addEventListener('keyup', function(e) {
             if (!document.getElementById('screen-page').classList.contains('active')) return;
             if (e.target.tagName === 'INPUT') return;
             if (e.key === 'F11' && !controlEnabled) return;
+            // 由 pasteToRemote 合成的完整 Ctrl+V 按下：物理 keyup 不再转发，避免 down/up 乱序
+            if (e.keyCode === 86 && skipNextCtrlVKeyup) {
+                skipNextCtrlVKeyup = false;
+                e.preventDefault();
+                return;
+            }
             e.preventDefault();
             sendKey(e.keyCode, false, e.altKey);
         });