diff --git a/.gitignore b/.gitignore
index f3e3a54..1cb9fa2 100644
--- a/.gitignore
+++ b/.gitignore
@@ -3,11 +3,13 @@ app/models/
 app/openvino-models/
 app/llm-models/
 app/speech-models/
+app/tts-models/
 
 # Downloaded portable runtimes/backends
 app/backend/
 app/llm-backend/
 app/speech-backend/
+app/tts-runtime/
 app/tools/
 
 # Generated files
@@ -15,6 +17,8 @@ app/config/
 app/dist/
 app/outputs/
 app/transcriptions/
+app/tts-outputs/
+app/tts-cache/
 
 # Dependencies and platform-specific copies
 **/node_modules/
diff --git a/app/frontend/src/App.jsx b/app/frontend/src/App.jsx
index 39b6607..2328c5e 100644
--- a/app/frontend/src/App.jsx
+++ b/app/frontend/src/App.jsx
@@ -6,7 +6,8 @@ import ModelManager from "./components/ModelManager";
 import Settings from "./components/Settings";
 import TextChat from "./components/TextChat";
 import SpeechTranscriber from "./components/SpeechTranscriber";
-import { cleanupCandidates, formatBytes, getCleanupCandidates, getDiagnostics, getHardwareSpecs, getHealth, getTelemetry, getBackendOptions, getBackendStatus, listGeneratedOutputs, listSpeechTranscriptions, deleteSpeechTranscription, stopServer } from "./services/api";
+import TextToSpeech from "./components/TextToSpeech";
+import { cleanupCandidates, formatBytes, getCleanupCandidates, getDiagnostics, getHardwareSpecs, getHealth, getTelemetry, getBackendOptions, getBackendStatus, listGeneratedOutputs, listSpeechTranscriptions, deleteSpeechTranscription, listTtsOutputs, deleteTtsOutput, stopServer } from "./services/api";
 import "./App.css";
 
 function App() {
@@ -183,6 +184,29 @@ function App() {
     localStorage.setItem("speechSettings", JSON.stringify(speechSettings));
   }, [speechSettings]);
 
+  const [ttsSettings, setTtsSettings] = useState(() => {
+    const saved = localStorage.getItem("ttsSettings");
+    if (saved) {
+      try {
+        const parsed = JSON.parse(saved);
+        return {
+          model: parsed.model || "",
+          voice: parsed.voice || "af_heart",
+          speed: Math.max(0.5, Math.min(2, Number(parsed.speed) || 1)),
+        };
+      } catch (_) {}
+    }
+    return {
+      model: "",
+      voice: "af_heart",
+      speed: 1,
+    };
+  });
+
+  useEffect(() => {
+    localStorage.setItem("ttsSettings", JSON.stringify(ttsSettings));
+  }, [ttsSettings]);
+
   // Lifted Chat History States
   const [conversations, setConversations] = useState([]);
   const [activeConversationId, setActiveConversationId] = useState(null);
@@ -190,6 +214,9 @@ function App() {
   const [speechTranscriptions, setSpeechTranscriptions] = useState([]);
   const [selectedSpeechTranscript, setSelectedSpeechTranscript] = useState(null);
   const [showSpeechHistory, setShowSpeechHistory] = useState(false);
+  const [ttsOutputs, setTtsOutputs] = useState([]);
+  const [selectedTtsOutput, setSelectedTtsOutput] = useState(null);
+  const [showTtsHistory, setShowTtsHistory] = useState(false);
 
   useEffect(() => {
     const saved = localStorage.getItem("chat_conversations");
@@ -315,6 +342,49 @@ function App() {
     refreshSpeechTranscriptions();
   }, [refreshSpeechTranscriptions]);
 
+  const refreshTtsOutputs = useCallback(async () => {
+    try {
+      const list = await listTtsOutputs();
+      setTtsOutputs(list);
+    } catch (err) {
+      console.warn("Could not load TTS outputs:", err);
+      setTtsOutputs([]);
+    }
+  }, []);
+
+  const handleDeleteTtsOutput = useCallback(async (item, e) => {
+    if (e) e.stopPropagation();
+    const itemId = item.filename || item.metadata;
+    if (!itemId) return;
+
+    const ok = await showConfirm({
+      title: "Delete TTS Output",
+      message: `Are you sure you want to delete "${item.displayName || item.audioFile || "this audio"}"? This action cannot be undone.`,
+      confirmLabel: "Delete",
+      danger: true
+    });
+    if (!ok) return;
+
+    try {
+      await deleteTtsOutput(itemId);
+      setSelectedTtsOutput((prev) => {
+        if (prev) {
+          const prevId = prev.filename || prev.metadata;
+          if (prevId === itemId) return null;
+        }
+        return prev;
+      });
+      refreshTtsOutputs();
+    } catch (err) {
+      console.error("Failed to delete TTS output:", err);
+      showAlert({ title: "Delete Failed", message: err.message || String(err), danger: true });
+    }
+  }, [showConfirm, showAlert, refreshTtsOutputs]);
+
+  useEffect(() => {
+    refreshTtsOutputs();
+  }, [refreshTtsOutputs]);
+
   // Load hardware specifications on mount
   useEffect(() => {
     async function loadSpecs() {
@@ -565,8 +635,14 @@ function App() {
       showSpeechHistory={showSpeechHistory}
       setShowSpeechHistory={setShowSpeechHistory}
       onDeleteSpeechTranscription={handleDeleteSpeechTranscription}
+      ttsOutputs={ttsOutputs}
+      selectedTtsOutput={selectedTtsOutput}
+      setSelectedTtsOutput={setSelectedTtsOutput}
+      showTtsHistory={showTtsHistory}
+      setShowTtsHistory={setShowTtsHistory}
+      onDeleteTtsOutput={handleDeleteTtsOutput}
     />
-  ), [activeTab, specs, conversations, activeConversationId, showHistory, handleDeleteConversation, speechTranscriptions, selectedSpeechTranscript, showSpeechHistory, handleDeleteSpeechTranscription]);
+  ), [activeTab, specs, conversations, activeConversationId, showHistory, handleDeleteConversation, speechTranscriptions, selectedSpeechTranscript, showSpeechHistory, handleDeleteSpeechTranscription, ttsOutputs, selectedTtsOutput, showTtsHistory, handleDeleteTtsOutput]);
 
   const handleStopServer = useCallback(async () => {
     if (!serverRunning || isStoppingServer) return;
@@ -673,6 +749,17 @@ function App() {
           />
         </div>
 
+        <div style={{ display: activeTab === "tts" ? "flex" : "none", flex: 1, flexDirection: "column", overflow: "hidden" }}>
+          <TextToSpeech
+            showAlert={showAlert}
+            showConfirm={showConfirm}
+            selectedOutput={selectedTtsOutput}
+            onOutputsChanged={refreshTtsOutputs}
+            ttsSettings={ttsSettings}
+            setTtsSettings={setTtsSettings}
+          />
+        </div>
+
         <div style={{ display: activeTab === "settings" ? "flex" : "none", flex: 1, flexDirection: "column", overflow: "hidden" }}>
           <Settings
             constraints={constraints}
@@ -689,6 +776,8 @@ function App() {
             setTextSettings={setTextSettings}
             speechSettings={speechSettings}
             setSpeechSettings={setSpeechSettings}
+            ttsSettings={ttsSettings}
+            setTtsSettings={setTtsSettings}
             health={health}
             cleanupItems={cleanupItems}
             isReadinessBusy={isReadinessBusy}
diff --git a/app/frontend/src/components/ModelManager.jsx b/app/frontend/src/components/ModelManager.jsx
index cc973a7..69d95e7 100644
--- a/app/frontend/src/components/ModelManager.jsx
+++ b/app/frontend/src/components/ModelManager.jsx
@@ -28,7 +28,14 @@ import {
   downloadSpeechModel,
   deleteSpeechModel,
   importSpeechModel,
-  getSpeechStatus
+  getSpeechStatus,
+  listTtsModels,
+  startTts,
+  stopTts,
+  downloadTtsModel,
+  deleteTtsModel,
+  importTtsModel,
+  getTtsStatus
 } from "../services/api";
 
 const MODEL_FILTERS = [
@@ -181,6 +188,38 @@ const OPENVINO_MODEL_LIBRARY = [
   },
 ];
 
+const TTS_MODEL_LIBRARY = [
+  {
+    group: "Kokoro Text-to-Speech Models",
+    items: [
+      {
+        id: "kokoro-onnx-q8",
+        name: "Kokoro 82M ONNX Q8",
+        filename: "kokoro-onnx-q8.json",
+        format: "Kokoro ONNX",
+        approxSize: "Model cache on first use",
+        resolution: "24 kHz WAV",
+        notes: "Recommended local TTS model. Good quality and faster startup than full precision on most PCs.",
+        url: "kokoro://install/kokoro-onnx-q8",
+        pageUrl: "https://huggingface.co/onnx-community/Kokoro-82M-v1.0-ONNX",
+        recommendedTiers: ["Low", "Mid", "High"],
+      },
+      {
+        id: "kokoro-onnx-fp32",
+        name: "Kokoro 82M ONNX FP32",
+        filename: "kokoro-onnx-fp32.json",
+        format: "Kokoro ONNX",
+        approxSize: "Model cache on first use",
+        resolution: "24 kHz WAV",
+        notes: "Full precision Kokoro ONNX path for quality testing. Uses more memory and disk cache.",
+        url: "kokoro://install/kokoro-onnx-fp32",
+        pageUrl: "https://huggingface.co/onnx-community/Kokoro-82M-v1.0-ONNX",
+        recommendedTiers: ["Mid", "High"],
+      },
+    ],
+  },
+];
+
 const COREML_MODEL_LIBRARY = [
   {
     group: "Apple Silicon NPU - CoreML Test",
@@ -227,7 +266,7 @@ function ModelManager({
   specs,
   textSettings,
 }) {
-  const [activeModelType, setActiveModelType] = useState("image"); // "image", "text", or "speech"
+  const [activeModelType, setActiveModelType] = useState("image"); // "image", "text", "speech", or "tts"
   const [localModels, setLocalModels] = useState([]);
   const [isLoadingModels, setIsLoadingModels] = useState(true);
   const [downloadingModelId, setDownloadingModelId] = useState(null);
@@ -248,6 +287,8 @@ function ModelManager({
   const [llmRunning, setLlmRunning] = useState(false);
   const [activeSpeechModel, setActiveSpeechModel] = useState(null);
   const [speechRunning, setSpeechRunning] = useState(false);
+  const [activeTtsModel, setActiveTtsModel] = useState(null);
+  const [ttsRunning, setTtsRunning] = useState(false);
   const [modelSearch, setModelSearch] = useState("");
   const [selectedFilters, setSelectedFilters] = useState([]);
   const [huggingFaceModels, setHuggingFaceModels] = useState([]);
@@ -271,13 +312,13 @@ function ModelManager({
     }
   };
 
-  const normalizedLocalModels = activeModelType === "speech"
+  const normalizedLocalModels = activeModelType === "speech" || activeModelType === "tts"
     ? localModels.filter((model) => model.filename)
     : localModels.map(normalizeModel).filter((model) => model.filename);
-  const displayedLocalModels = activeModelType === "speech"
+  const displayedLocalModels = activeModelType === "speech" || activeModelType === "tts"
     ? normalizedLocalModels.filter((model) => model.installed)
     : normalizedLocalModels.filter((model) => !model.isProjector);
-  const allModelNames = activeModelType === "speech"
+  const allModelNames = activeModelType === "speech" || activeModelType === "tts"
     ? displayedLocalModels.map((model) => model.filename)
     : normalizedLocalModels.map((model) => model.filename);
   const modelNames = displayedLocalModels.map((model) => model.filename);
@@ -306,7 +347,7 @@ function ModelManager({
       group: "Recommended Text Models from Hugging Face",
       items: hasHuggingFaceResults ? displayedHuggingFaceModels : TEXT_MODEL_LIBRARY[0].items,
     }];
-  } else {
+  } else if (activeModelType === "speech") {
     visibleModelLibrary = [{
       group: "Whisper Speech Models",
       items: normalizedLocalModels.filter((model) => model.url).map((model) => ({
@@ -317,6 +358,8 @@ function ModelManager({
         recommendedTiers: model.recommended ? ["Low", "Mid", "High"] : [],
       })),
     }];
+  } else {
+    visibleModelLibrary = TTS_MODEL_LIBRARY;
   }
 
   useEffect(() => {
@@ -393,9 +436,12 @@ function ModelManager({
       } else if (activeModelType === "text") {
         const list = await listLlmModels();
         setLocalModels(list.map(normalizeModel).filter((model) => model.filename));
-      } else {
+      } else if (activeModelType === "speech") {
         const list = await listSpeechModels();
         setLocalModels(list.filter((model) => model.filename));
+      } else {
+        const list = await listTtsModels();
+        setLocalModels(list.filter((model) => model.filename));
       }
     } catch (e) {
       console.error(e);
@@ -421,10 +467,11 @@ function ModelManager({
     let cancelled = false;
     const updateBackendInfo = async () => {
       try {
-        const [sdStatus, llmStatus, speechStatus] = await Promise.all([
+        const [sdStatus, llmStatus, speechStatus, ttsStatus] = await Promise.all([
           getBackendStatus(),
           getLlmStatus(),
-          getSpeechStatus()
+          getSpeechStatus(),
+          getTtsStatus()
         ]);
         if (cancelled) return;
         
@@ -434,6 +481,7 @@ function ModelManager({
           backendDevice: sdStatus.settings?.backendDevice || sdStatus.loading?.device || "",
           llmBackendMode: llmStatus.settings?.backendMode || llmStatus.settings?.backendBinary || "",
           speechBackendMode: speechStatus.settings?.backendMode || speechStatus.backendMode || "",
+          ttsBackendMode: ttsStatus.settings?.backendMode || ttsStatus.backendMode || "",
         });
         
         if (llmStatus.ready) {
@@ -450,6 +498,13 @@ function ModelManager({
           setActiveSpeechModel(null);
           setSpeechRunning(false);
         }
+        if (ttsStatus.ready) {
+          setActiveTtsModel(ttsStatus.settings?.model || null);
+          setTtsRunning(true);
+        } else {
+          setActiveTtsModel(null);
+          setTtsRunning(false);
+        }
       } catch (_) {}
     };
 
@@ -500,8 +555,10 @@ function ModelManager({
         res = await downloadModel(url);
       } else if (activeModelType === "text") {
         res = await downloadLlmModel(url, expectedFilename, companion);
-      } else {
+      } else if (activeModelType === "speech") {
         res = await downloadSpeechModel(url, expectedFilename);
+      } else {
+        res = await downloadTtsModel(url, expectedFilename);
       }
       if (res && res.ok) {
         if (!companion && res.projectorUrl && res.projectorFilename) {
@@ -563,7 +620,7 @@ function ModelManager({
     e.preventDefault();
     if (!downloadUrl.trim()) return;
 
-    let modelName = activeModelType === "image" ? "model.safetensors" : activeModelType === "text" ? "model.gguf" : "model.bin";
+    let modelName = activeModelType === "image" ? "model.safetensors" : activeModelType === "text" ? "model.gguf" : activeModelType === "speech" ? "model.bin" : "model.json";
     try {
       const parsed = new URL(downloadUrl.trim());
       modelName = parsed.pathname.split("/").pop() || modelName;
@@ -591,6 +648,18 @@ function ModelManager({
       return;
     }
 
+    if (activeModelType === "tts") {
+      const confirmed = await showConfirm({
+        title: "Download TTS Model?",
+        message: `Install "${model.name}"? Kokoro model files are cached on first generation.`,
+        confirmLabel: "Install",
+      });
+      if (confirmed) {
+        await downloadByUrl(model.id || model.url, model.filename);
+      }
+      return;
+    }
+
     if (model.backendType === "openvino-npu") {
       const confirmed = await showConfirm({
         title: "Download OpenVINO Model?",
@@ -734,7 +803,7 @@ function ModelManager({
         setModelLoadProgress(null);
         setLoadingModelId(null);
       }
-    } else {
+    } else if (activeModelType === "speech") {
       setLoadingModelId(modelId);
       setModelLoadProgress({
         progress: 30,
@@ -763,6 +832,38 @@ function ModelManager({
         setModelLoadProgress(null);
         setLoadingModelId(null);
       }
+    } else {
+      setLoadingModelId(modelId);
+      setModelLoadProgress({
+        progress: 30,
+        phase: "Starting Kokoro ONNX TTS runtime...",
+        speed: "",
+        current: 0,
+        total: 0,
+        model: modelId,
+        backendMode: "",
+        backendBinary: "",
+        device: "",
+      });
+      try {
+        await stopServer();
+        await stopLlm();
+        await stopSpeech();
+        await startTts(modelId);
+        setActiveTtsModel(modelId);
+        setTtsRunning(true);
+        setActiveSpeechModel(null);
+        setSpeechRunning(false);
+        setActiveLlmModel(null);
+        setLlmRunning(false);
+        setActiveModel(null);
+        setServerRunning(false);
+      } catch (err) {
+        showAlert({ title: "TTS Model Load Failed", message: err.message || String(err), danger: true });
+      } finally {
+        setModelLoadProgress(null);
+        setLoadingModelId(null);
+      }
     }
   };
 
@@ -966,7 +1067,7 @@ function ModelManager({
         setIsUnloading(false);
         setUnloadProgress({ progress: 0, phase: "" });
       }
-    } else {
+    } else if (activeModelType === "speech") {
       setIsUnloading(true);
       setUnloadProgress({ progress: 50, phase: "Stopping whisper.cpp speech runtime..." });
       try {
@@ -979,6 +1080,19 @@ function ModelManager({
         setIsUnloading(false);
         setUnloadProgress({ progress: 0, phase: "" });
       }
+    } else {
+      setIsUnloading(true);
+      setUnloadProgress({ progress: 50, phase: "Stopping Kokoro TTS runtime..." });
+      try {
+        await stopTts();
+        setActiveTtsModel(null);
+        setTtsRunning(false);
+      } catch (e) {
+        showAlert({ title: "Unload Failed", message: e.message || String(e), danger: true });
+      } finally {
+        setIsUnloading(false);
+        setUnloadProgress({ progress: 0, phase: "" });
+      }
     }
   };
 
@@ -1000,11 +1114,16 @@ function ModelManager({
           if (activeLlmModel === filename) {
             await handleUnloadModel();
           }
-        } else {
+        } else if (activeModelType === "speech") {
           await deleteSpeechModel(filename);
           if (activeSpeechModel === filename) {
             await handleUnloadModel();
           }
+        } else {
+          await deleteTtsModel(filename);
+          if (activeTtsModel === filename) {
+            await handleUnloadModel();
+          }
         }
         await fetchModels();
       } catch (err) {
@@ -1046,7 +1165,7 @@ function ModelManager({
             status: progressData.status
           });
         }, controller.signal);
-      } else {
+      } else if (activeModelType === "speech") {
         await importSpeechModel(sourcePath, (progressData) => {
           setImportProgress(Math.round(progressData.progress));
           setImportInfo({
@@ -1056,6 +1175,16 @@ function ModelManager({
             status: progressData.status
           });
         }, controller.signal);
+      } else {
+        await importTtsModel(sourcePath, (progressData) => {
+          setImportProgress(Math.round(progressData.progress));
+          setImportInfo({
+            filename: progressData.filename,
+            speed: progressData.speed_mb_s.toFixed(1),
+            eta: Math.round(progressData.eta_secs),
+            status: progressData.status
+          });
+        }, controller.signal);
       }
 
       await fetchModels();
@@ -1112,6 +1241,13 @@ function ModelManager({
         >
           Speech Models (Whisper)
         </button>
+        <button
+          className={`m3-btn ${activeModelType === "tts" ? "m3-btn-filled" : "m3-btn-outlined"}`}
+          onClick={() => setActiveModelType("tts")}
+          style={{ height: "40px", padding: "0 20px" }}
+        >
+          TTS Models (Kokoro)
+        </button>
       </div>
 
 
@@ -1190,6 +1326,26 @@ function ModelManager({
         </div>
       )}
 
+      {activeModelType === "tts" && activeTtsModel && (
+        <div className="m3-card" style={{ borderLeft: "4px solid var(--md-sys-color-primary)", background: "var(--md-sys-color-primary-container)", color: "var(--md-sys-color-on-primary-container)" }}>
+          <div style={{ display: "flex", justifyContent: "space-between", alignItems: "center" }}>
+            <div>
+              <h4 style={{ fontWeight: 700 }}>Active TTS Model: {activeTtsModel}</h4>
+              <p style={{ fontSize: "0.85rem", marginTop: "2px", opacity: 0.9 }}>
+                The local Kokoro ONNX runtime is ready. Text to Speech can generate WAV narration.
+              </p>
+              {backendInfo.ttsBackendMode && (
+                <div style={{ display: "flex", gap: "8px", marginTop: "8px", flexWrap: "wrap" }}>
+                  <span className="status-chip" style={{ cursor: "default", background: "rgba(255,255,255,0.15)", color: "inherit", border: "1px solid rgba(255,255,255,0.2)" }}>
+                    <span>{backendInfo.ttsBackendMode}</span>
+                  </span>
+                </div>
+              )}
+            </div>
+          </div>
+        </div>
+      )}
+
       {/* Loading Progress Bar */}
       {modelLoadProgress && (
         <div className="m3-card" style={{ borderLeft: "4px solid var(--md-sys-color-primary)", marginTop: "24px" }}>
@@ -1241,7 +1397,7 @@ function ModelManager({
       <div className="m3-card" style={{ marginTop: "24px" }}>
         <h3 className="m3-card-title">
           <Database size={18} style={{ color: "var(--md-sys-color-primary)" }} />
-          Local {activeModelType === "image" ? "Image" : activeModelType === "text" ? "Text" : "Speech"} Models ({displayedLocalModels.length})
+          Local {activeModelType === "image" ? "Image" : activeModelType === "text" ? "Text" : activeModelType === "speech" ? "Speech" : "TTS"} Models ({displayedLocalModels.length})
         </h3>
         
         {isLoadingModels ? (
@@ -1255,13 +1411,15 @@ function ModelManager({
               ? "No image models detected in app/models/. Download from the library below or import a file."
               : activeModelType === "text"
               ? "No text models detected in app/llm-models/. Download from the library below or import a file."
-              : "No speech models detected in app/speech-models/. Download from the library below or import a .bin file."}
+              : activeModelType === "speech"
+              ? "No speech models detected in app/speech-models/. Download from the library below or import a .bin file."
+              : "No TTS models detected in app/tts-models/. Install Kokoro from the library below or import a .json manifest."}
           </p>
         ) : (
           <div style={{ display: "flex", flexDirection: "column", gap: "10px" }}>
             {displayedLocalModels.map((model) => {
               const filename = model.filename;
-              const isActive = activeModelType === "image" ? activeModel === filename : activeModelType === "text" ? activeLlmModel === filename : activeSpeechModel === filename;
+              const isActive = activeModelType === "image" ? activeModel === filename : activeModelType === "text" ? activeLlmModel === filename : activeModelType === "speech" ? activeSpeechModel === filename : activeTtsModel === filename;
               
               return (
                 <div 
@@ -1284,7 +1442,9 @@ function ModelManager({
                         ? (model.backendType === "openvino-npu" ? "OpenVINO NPU Model" : model.format || "Local Weights File")
                         : activeModelType === "text"
                         ? "llama.cpp GGUF Model"
-                        : `${model.language || "Whisper"} Model`
+                        : activeModelType === "speech"
+                        ? `${model.language || "Whisper"} Model`
+                        : "Kokoro ONNX Model"
                       } • {model.size || formatBytes(model.sizeBytes)}
                     </span>
                   </div>
@@ -1332,7 +1492,7 @@ function ModelManager({
       <div className="workspace-title-section" style={{ marginTop: "32px", marginBottom: "16px" }}>
         <h3 className="m3-card-title">
           <Library size={20} style={{ color: "var(--md-sys-color-primary)" }} />
-          {activeModelType === "text" ? "Hugging Face Model Library" : activeModelType === "speech" ? "Speech Model Library" : "Model Library"}
+          {activeModelType === "text" ? "Hugging Face Model Library" : activeModelType === "speech" ? "Speech Model Library" : activeModelType === "tts" ? "TTS Model Library" : "Model Library"}
         </h3>
       </div>
 
@@ -1541,14 +1701,14 @@ function ModelManager({
           <input
             type="file"
             style={{ display: "none" }}
-            accept={activeModelType === "image" ? ".safetensors,.ckpt" : activeModelType === "text" ? ".gguf" : ".bin"}
+            accept={activeModelType === "image" ? ".safetensors,.ckpt" : activeModelType === "text" ? ".gguf" : activeModelType === "speech" ? ".bin" : ".json"}
             onChange={handleImportFile}
             disabled={importProgress !== null}
           />
           <FolderOpen className="import-icon" />
           <span style={{ fontWeight: 600 }}>Choose weights file</span>
           <span style={{ fontSize: "0.75rem", color: "var(--md-sys-color-outline)", textAlign: "center" }}>
-            Select {activeModelType === "image" ? "`.safetensors` or `.ckpt` weights." : activeModelType === "text" ? "`.gguf` weights." : "`whisper.cpp .bin` weights."}
+            Select {activeModelType === "image" ? "`.safetensors` or `.ckpt` weights." : activeModelType === "text" ? "`.gguf` weights." : activeModelType === "speech" ? "`whisper.cpp .bin` weights." : "`Kokoro .json` manifest."}
           </span>
         </label>
 
@@ -1558,7 +1718,7 @@ function ModelManager({
             Download Model from URL
           </h4>
           <p style={{ fontSize: "0.75rem", color: "var(--md-sys-color-outline)", marginBottom: "12px" }}>
-            Download any {activeModelType === "image" ? "Safetensors" : activeModelType === "text" ? "GGUF" : "Whisper .bin"} model from Hugging Face directly to your models folder.
+            Download any {activeModelType === "image" ? "Safetensors" : activeModelType === "text" ? "GGUF" : activeModelType === "speech" ? "Whisper .bin" : "Kokoro manifest"} model from Hugging Face directly to your models folder.
           </p>
           {isUrlDownloading ? (
             <div className="model-progress-section" style={{ marginTop: "0px" }}>
diff --git a/app/frontend/src/components/Settings.jsx b/app/frontend/src/components/Settings.jsx
index 0597fac..cecbddf 100644
--- a/app/frontend/src/components/Settings.jsx
+++ b/app/frontend/src/components/Settings.jsx
@@ -207,6 +207,8 @@ function Settings({
   setTextSettings,
   speechSettings,
   setSpeechSettings,
+  ttsSettings,
+  setTtsSettings,
   showAlert = async ({ message }) => window.alert(message),
   showConfirm = async ({ message }) => window.confirm(message),
   health,
@@ -230,6 +232,7 @@ function Settings({
       image: localStorage.getItem("settings_section_image") === "true",
       text: localStorage.getItem("settings_section_text") === "true",
       speech: localStorage.getItem("settings_section_speech") === "true",
+      tts: localStorage.getItem("settings_section_tts") === "true",
     };
   });
 
@@ -297,6 +300,13 @@ function Settings({
     }));
   };
 
+  const updateTtsSetting = (key, value) => {
+    setTtsSettings((prev) => ({
+      ...prev,
+      [key]: value
+    }));
+  };
+
   const buildTextStartOptions = (settings) => ({
     threads: settings?.threads || specs?.cpu_cores_physical || 4,
     contextSize: settings?.contextSize ?? 0,
@@ -1037,6 +1047,93 @@ function Settings({
   };
 
   // ─── Appearance Settings ───
+  const renderTtsSettings = () => {
+    const TTS_VOICES = [
+      { value: "af_heart", label: "Heart - Female, US English" },
+      { value: "af_bella", label: "Bella - Female, US English" },
+      { value: "af_nicole", label: "Nicole - Female, US English" },
+      { value: "af_sarah", label: "Sarah - Female, US English" },
+      { value: "am_michael", label: "Michael - Male, US English" },
+      { value: "am_fenrir", label: "Fenrir - Male, US English" },
+      { value: "bf_emma", label: "Emma - Female, UK English" },
+      { value: "bm_george", label: "George - Male, UK English" },
+    ];
+
+    return (
+      <>
+        <SectionHeader
+          icon={Volume2}
+          title="Text to Speech"
+          count={2}
+          color="#8b5cf6"
+          isExpanded={expandedSections.tts}
+          onToggle={() => toggleSection("tts")}
+        />
+
+        {expandedSections.tts && (
+          <div className="settings-expanded-content">
+            <div className="settings-two-column">
+              <div className="settings-column">
+                <div className="settings-subsection">
+                  <div className="settings-subsection-title">
+                    <Volume2 size={16} />
+                    Voice Defaults
+                  </div>
+                  <div className="m3-field-group">
+                    <div className="m3-text-field">
+                      <label className="m3-text-field-label">Default Voice</label>
+                      <select
+                        className="m3-input"
+                        value={ttsSettings?.voice || "af_heart"}
+                        onChange={(e) => updateTtsSetting("voice", e.target.value)}
+                      >
+                        {TTS_VOICES.map((item) => (
+                          <option key={item.value} value={item.value}>
+                            {item.label}
+                          </option>
+                        ))}
+                      </select>
+                      <span className="settings-option-desc" style={{ marginTop: "4px", display: "block" }}>
+                        Default Kokoro voice for generated WAV files.
+                      </span>
+                    </div>
+                  </div>
+                </div>
+              </div>
+
+              <div className="settings-column">
+                <div className="settings-subsection">
+                  <div className="settings-subsection-title">
+                    <Gauge size={16} />
+                    Speech Timing
+                  </div>
+                  <div className="m3-slider-group">
+                    <div className="m3-slider-header">
+                      <span className="m3-slider-label">Speed</span>
+                      <span className="settings-value-badge">{(ttsSettings?.speed || 1).toFixed(2)}x</span>
+                    </div>
+                    <input
+                      type="range"
+                      className="m3-slider"
+                      value={ttsSettings?.speed || 1}
+                      onChange={(e) => updateTtsSetting("speed", parseFloat(e.target.value))}
+                      min="0.5"
+                      max="2"
+                      step="0.05"
+                    />
+                    <span className="settings-option-desc">
+                      1.00x is natural speed. Lower values are slower, higher values are faster.
+                    </span>
+                  </div>
+                </div>
+              </div>
+            </div>
+          </div>
+        )}
+      </>
+    );
+  };
+
   const renderAppearanceSettings = () => (
     <>
       <SectionHeader 
@@ -1144,6 +1241,9 @@ function Settings({
 
       {/* Speech Settings Section */}
       {renderSpeechSettings()}
+
+      {/* Text to Speech Settings Section */}
+      {renderTtsSettings()}
     </div>
   );
 }
diff --git a/app/frontend/src/components/Sidebar.jsx b/app/frontend/src/components/Sidebar.jsx
index 5dbf93f..b3b60b3 100644
--- a/app/frontend/src/components/Sidebar.jsx
+++ b/app/frontend/src/components/Sidebar.jsx
@@ -1,5 +1,5 @@
 import React, { memo } from "react";
-import { Image, FolderDown, MessageSquare, Mic, Settings, Shield, Terminal, ChevronDown, ChevronUp, Trash2 } from "lucide-react";
+import { Image, FolderDown, MessageSquare, Mic, Settings, Shield, Terminal, ChevronDown, ChevronUp, Trash2, Volume2 } from "lucide-react";
 
 function formatSidebarDate(value) {
   const date = new Date(value);
@@ -24,7 +24,13 @@ function Sidebar({
   setSelectedSpeechTranscript,
   showSpeechHistory,
   setShowSpeechHistory,
-  onDeleteSpeechTranscription
+  onDeleteSpeechTranscription,
+  ttsOutputs = [],
+  selectedTtsOutput,
+  setSelectedTtsOutput,
+  showTtsHistory,
+  setShowTtsHistory,
+  onDeleteTtsOutput
 }) {
   return (
     <div className="sidebar">
@@ -283,6 +289,127 @@ function Sidebar({
             )}
           </div>
 
+          <div className="nav-item-wrapper" style={{ display: "flex", flexDirection: "column" }}>
+            <div
+              className={`nav-item ${activeTab === "tts" ? "active" : ""}`}
+              onClick={() => setActiveTab("tts")}
+              style={{ display: "flex", justifyContent: "space-between", alignItems: "center", width: "100%", boxSizing: "border-box" }}
+            >
+              <div style={{ display: "flex", alignItems: "center", gap: "12px" }}>
+                <Volume2 size={20} />
+                <span>Text to Speech</span>
+              </div>
+              <button
+                onClick={(e) => {
+                  e.stopPropagation();
+                  setShowTtsHistory(!showTtsHistory);
+                }}
+                style={{
+                  background: "none",
+                  border: "none",
+                  color: "inherit",
+                  cursor: "pointer",
+                  padding: "4px",
+                  display: "flex",
+                  alignItems: "center",
+                  justifyContent: "center",
+                  borderRadius: "50%",
+                  transition: "background-color 0.2s"
+                }}
+                className="history-toggle-arrow"
+                title={showTtsHistory ? "Hide TTS Outputs" : "Show TTS Outputs"}
+              >
+                {showTtsHistory ? <ChevronUp size={16} /> : <ChevronDown size={16} />}
+              </button>
+            </div>
+
+            {showTtsHistory && (
+              <div
+                className="sidebar-history-list"
+                style={{
+                  paddingLeft: "14px",
+                  marginTop: "6px",
+                  marginBottom: "6px",
+                  display: "flex",
+                  flexDirection: "column",
+                  gap: "4px",
+                  maxHeight: "220px",
+                  overflowY: "auto",
+                  borderLeft: "2px solid var(--border-color)"
+                }}
+              >
+                {ttsOutputs.length === 0 ? (
+                  <div style={{ padding: "8px 12px", fontSize: "0.78rem", color: "var(--md-sys-color-outline)", opacity: 0.8 }}>
+                    No saved audio
+                  </div>
+                ) : (
+                  ttsOutputs.map((item) => {
+                    const itemId = item.filename || item.metadata || item.audioFile;
+                    const isActive = selectedTtsOutput && (selectedTtsOutput.filename || selectedTtsOutput.metadata || selectedTtsOutput.audioFile) === itemId;
+                    const title = item.displayName || item.text || item.audioFile || item.filename || "TTS Output";
+                    const date = formatSidebarDate(item.modifiedAt || item.createdAt);
+                    return (
+                      <div
+                        key={itemId}
+                        onClick={() => {
+                          setSelectedTtsOutput(item);
+                          setActiveTab("tts");
+                        }}
+                        style={{
+                          display: "flex",
+                          flexDirection: "row",
+                          alignItems: "center",
+                          justifyContent: "space-between",
+                          padding: "7px 8px 7px 10px",
+                          borderRadius: "var(--md-shape-corner-small)",
+                          fontSize: "0.78rem",
+                          cursor: "pointer",
+                          background: isActive ? "var(--md-sys-color-secondary-container)" : "transparent",
+                          color: isActive ? "var(--md-sys-color-on-secondary-container)" : "var(--md-sys-color-on-surface-variant)",
+                          border: isActive ? "1px solid var(--md-sys-color-outline-variant)" : "1px solid transparent",
+                          transition: "background 0.2s"
+                        }}
+                        className="sidebar-history-item"
+                        title={title}
+                      >
+                        <div style={{ display: "flex", flexDirection: "column", gap: "2px", flex: 1, minWidth: 0 }}>
+                          <span style={{ overflow: "hidden", textOverflow: "ellipsis", whiteSpace: "nowrap", fontWeight: isActive ? 600 : 400 }}>
+                            {title}
+                          </span>
+                          {date && (
+                            <span style={{ fontSize: "0.68rem", opacity: 0.72, overflow: "hidden", textOverflow: "ellipsis", whiteSpace: "nowrap" }}>
+                              {date}
+                            </span>
+                          )}
+                        </div>
+                        <button
+                          onClick={(e) => {
+                            e.stopPropagation();
+                            onDeleteTtsOutput?.(item, e);
+                          }}
+                          style={{
+                            background: "none",
+                            border: "none",
+                            color: "var(--md-sys-color-outline)",
+                            cursor: "pointer",
+                            padding: "2px",
+                            marginLeft: "6px",
+                            display: "flex",
+                            alignItems: "center"
+                          }}
+                          className="sidebar-history-delete"
+                          title="Delete TTS Output"
+                        >
+                          <Trash2 size={12} />
+                        </button>
+                      </div>
+                    );
+                  })
+                )}
+              </div>
+            )}
+          </div>
+
           <div
             className={`nav-item ${activeTab === "models" ? "active" : ""}`}
             onClick={() => setActiveTab("models")}
diff --git a/app/frontend/src/components/TextToSpeech.jsx b/app/frontend/src/components/TextToSpeech.jsx
new file mode 100644
index 0000000..a621eda
--- /dev/null
+++ b/app/frontend/src/components/TextToSpeech.jsx
@@ -0,0 +1,317 @@
+import React, { useCallback, useEffect, useMemo, useRef, useState } from "react";
+import {
+  LoaderCircle,
+  Play,
+  RefreshCw,
+  Save,
+  Square,
+  Volume2,
+} from "lucide-react";
+import {
+  getTtsStatus,
+  listTtsModels,
+  speakTts,
+  startTts,
+  stopTts,
+} from "../services/api";
+
+const FALLBACK_VOICES = [
+  { id: "af_heart", name: "Heart", language: "en-us", gender: "Female" },
+  { id: "af_bella", name: "Bella", language: "en-us", gender: "Female" },
+  { id: "am_michael", name: "Michael", language: "en-us", gender: "Male" },
+  { id: "bf_emma", name: "Emma", language: "en-gb", gender: "Female" },
+  { id: "bm_george", name: "George", language: "en-gb", gender: "Male" },
+];
+
+function downloadUrl(url, filename) {
+  const link = document.createElement("a");
+  link.href = url;
+  link.download = filename;
+  document.body.appendChild(link);
+  link.click();
+  link.remove();
+}
+
+export default function TextToSpeech({
+  showAlert,
+  selectedOutput,
+  onOutputsChanged,
+  ttsSettings,
+  setTtsSettings,
+}) {
+  const [status, setStatus] = useState({ ready: false, running: false, runtimeInstalled: false, settings: {}, voices: [] });
+  const [models, setModels] = useState([]);
+  const [selectedModel, setSelectedModel] = useState("");
+  const [text, setText] = useState("Hello from Local AI Studio text to speech.");
+  const [isLoadingModel, setIsLoadingModel] = useState(false);
+  const [isGenerating, setIsGenerating] = useState(false);
+  const [output, setOutput] = useState(null);
+  const abortRef = useRef(null);
+
+  const installedModels = useMemo(() => models.filter((model) => model.installed), [models]);
+  const voices = status.voices?.length ? status.voices : FALLBACK_VOICES;
+  const selectedVoice = ttsSettings?.voice || "af_heart";
+  const speed = ttsSettings?.speed || 1;
+
+  const updateTtsSetting = (key, value) => {
+    setTtsSettings((prev) => ({ ...prev, [key]: value }));
+  };
+
+  const refresh = useCallback(async () => {
+    const [nextStatus, nextModels] = await Promise.all([
+      getTtsStatus(),
+      listTtsModels(),
+    ]);
+    setStatus(nextStatus);
+    setModels(nextModels);
+    setSelectedModel((current) => {
+      const installed = nextModels.filter((model) => model.installed);
+      if (current && installed.some((model) => model.filename === current || model.id === current)) return current;
+      const active = nextStatus.settings?.model;
+      if (active && installed.some((model) => model.filename === active)) return active;
+      const preferred = ttsSettings?.model;
+      if (preferred && installed.some((model) => model.filename === preferred)) return preferred;
+      return installed.find((model) => model.recommended)?.filename || installed[0]?.filename || "";
+    });
+  }, [ttsSettings?.model]);
+
+  useEffect(() => {
+    refresh().catch((err) => showAlert?.({ title: "TTS Status Failed", message: err.message || String(err), danger: true }));
+    const interval = setInterval(() => {
+      refresh().catch(() => {});
+    }, 2500);
+    return () => clearInterval(interval);
+  }, [refresh, showAlert]);
+
+  useEffect(() => {
+    if (!selectedOutput) return;
+    setOutput(selectedOutput);
+    setText(selectedOutput.text || "");
+    if (selectedOutput.voice) updateTtsSetting("voice", selectedOutput.voice);
+    if (selectedOutput.speed) updateTtsSetting("speed", selectedOutput.speed);
+    if (selectedOutput.model) setSelectedModel(selectedOutput.model);
+  }, [selectedOutput]);
+
+  useEffect(() => {
+    return () => abortRef.current?.abort();
+  }, []);
+
+  const handleLoadModel = async () => {
+    if (!selectedModel) return;
+    setIsLoadingModel(true);
+    try {
+      await startTts(selectedModel, { voice: selectedVoice, speed });
+      updateTtsSetting("model", selectedModel);
+      await refresh();
+    } catch (err) {
+      showAlert?.({ title: "TTS Model Load Failed", message: err.message || String(err), danger: true });
+    } finally {
+      setIsLoadingModel(false);
+    }
+  };
+
+  const handleStop = async () => {
+    try {
+      await stopTts();
+      await refresh();
+    } catch (err) {
+      showAlert?.({ title: "Stop TTS Failed", message: err.message || String(err), danger: true });
+    }
+  };
+
+  const handleGenerate = async () => {
+    if (!text.trim() || !selectedModel) return;
+    setIsGenerating(true);
+    const controller = new AbortController();
+    abortRef.current = controller;
+    try {
+      if (!status.ready || status.settings?.model !== selectedModel) {
+        setIsLoadingModel(true);
+        try {
+          await startTts(selectedModel, { voice: selectedVoice, speed });
+          updateTtsSetting("model", selectedModel);
+          await refresh();
+        } finally {
+          setIsLoadingModel(false);
+        }
+      }
+      const generated = await speakTts(text, {
+        model: selectedModel,
+        voice: selectedVoice,
+        speed,
+        signal: controller.signal,
+      });
+      setOutput(generated);
+      await refresh();
+      await onOutputsChanged?.();
+    } catch (err) {
+      if (err.name !== "AbortError") {
+        showAlert?.({ title: "TTS Generation Failed", message: err.message || String(err), danger: true });
+      }
+    } finally {
+      abortRef.current = null;
+      setIsGenerating(false);
+    }
+  };
+
+  const handleCancel = () => {
+    abortRef.current?.abort();
+    setIsGenerating(false);
+  };
+
+  const isLoaded = status.ready && status.settings?.model === selectedModel;
+
+  return (
+    <div className="workspace-area speech-workspace">
+      <div className="workspace-title-section">
+        <h2 className="workspace-title">Text to Speech</h2>
+        <p className="workspace-subtitle">
+          Generate local WAV narration with Kokoro ONNX.
+        </p>
+      </div>
+
+      <div className="speech-grid">
+        <section className="m3-card speech-panel">
+          <div className="speech-panel-header">
+            <h3>Runtime</h3>
+            <span className={`status-chip ${status.runtimeInstalled ? "" : "offline"}`}>
+              {status.runtimeInstalled ? (status.ready ? "Ready" : "Installed") : "Runtime missing"}
+            </span>
+          </div>
+
+          {status.error && <div className="text-progress error">{status.error}</div>}
+
+          <label className="speech-label">
+            Kokoro model
+            <select className="m3-input" value={selectedModel} onChange={(event) => setSelectedModel(event.target.value)}>
+              {installedModels.length === 0 && <option value="">No downloaded TTS models</option>}
+              {installedModels.map((model) => (
+                <option key={model.id || model.filename} value={model.filename}>
+                  {model.name || model.filename}
+                </option>
+              ))}
+            </select>
+          </label>
+
+          <div className="speech-controls-row">
+            <label className="speech-label">
+              Voice
+              <select className="m3-input" value={selectedVoice} onChange={(event) => updateTtsSetting("voice", event.target.value)}>
+                {voices.map((voice) => (
+                  <option key={voice.id} value={voice.id}>
+                    {voice.name} ({voice.gender}, {voice.language})
+                  </option>
+                ))}
+              </select>
+            </label>
+            <label className="speech-label">
+              Speed
+              <input
+                className="m3-input"
+                type="number"
+                min="0.5"
+                max="2"
+                step="0.05"
+                value={speed}
+                onChange={(event) => updateTtsSetting("speed", Math.max(0.5, Math.min(2, Number(event.target.value) || 1)))}
+              />
+            </label>
+          </div>
+
+          <div className="speech-button-row">
+            <button
+              className="m3-btn m3-btn-filled"
+              onClick={handleLoadModel}
+              disabled={!selectedModel || isLoadingModel || isGenerating || isLoaded}
+            >
+              {isLoadingModel ? <LoaderCircle className="progress-spinner" size={14} /> : <Play size={14} />}
+              <span>{isLoaded ? "Loaded" : isLoadingModel ? "Loading" : "Load"}</span>
+            </button>
+            <button className="m3-btn m3-btn-outlined" onClick={refresh}>
+              <RefreshCw size={14} />
+              <span>Refresh</span>
+            </button>
+            {status.ready && (
+              <button className="m3-btn m3-btn-error" onClick={handleStop}>
+                <Square size={14} />
+                <span>Stop Runtime</span>
+              </button>
+            )}
+          </div>
+
+          {installedModels.length === 0 && (
+            <div className="text-progress">
+              Download a Kokoro model from Model Manager, TTS Models first.
+            </div>
+          )}
+        </section>
+
+        <section className="m3-card speech-panel">
+          <div className="speech-panel-header">
+            <h3>Input</h3>
+            <span className="status-chip">{text.trim().length} chars</span>
+          </div>
+          <textarea
+            className="m3-input"
+            value={text}
+            onChange={(event) => setText(event.target.value)}
+            rows={10}
+            maxLength={5000}
+            style={{ width: "100%", resize: "vertical", minHeight: "220px", lineHeight: 1.45 }}
+            placeholder="Type text to turn into speech..."
+          />
+          <div className="speech-button-row">
+            {isGenerating ? (
+              <button className="m3-btn m3-btn-error" onClick={handleCancel}>
+                <Square size={14} />
+                <span>Cancel</span>
+              </button>
+            ) : (
+              <button
+                className="m3-btn m3-btn-filled"
+                onClick={handleGenerate}
+                disabled={!text.trim() || !selectedModel || !status.runtimeInstalled}
+              >
+                <Volume2 size={14} />
+                <span>Generate WAV</span>
+              </button>
+            )}
+            {isGenerating && <span className="text-progress"><LoaderCircle className="progress-spinner" size={14} /> Generating locally...</span>}
+          </div>
+        </section>
+      </div>
+
+      <section className="m3-card speech-result-panel">
+        <div className="speech-panel-header">
+          <h3>Output</h3>
+          <div className="speech-button-row">
+            <button className="m3-btn m3-btn-outlined" onClick={() => output?.url && downloadUrl(output.url, output.audioFile || "tts.wav")} disabled={!output?.url}>
+              <Save size={14} />
+              <span>WAV</span>
+            </button>
+          </div>
+        </div>
+        {output?.url ? (
+          <div style={{ display: "flex", flexDirection: "column", gap: "14px" }}>
+            <audio controls src={output.url} style={{ width: "100%" }} />
+            <div className="text-progress">
+              {output.voiceName || output.voice} &bull; {output.modelName || output.model} &bull; {new Date(output.createdAt).toLocaleString()}
+            </div>
+            <textarea
+              className="m3-input"
+              value={output.text || text}
+              onChange={(event) => setOutput((prev) => ({ ...(prev || {}), text: event.target.value }))}
+              rows={5}
+              style={{ width: "100%", resize: "vertical", minHeight: "120px", lineHeight: 1.45 }}
+            />
+          </div>
+        ) : (
+          <div className="empty-state" style={{ padding: "32px 0" }}>
+            <Volume2 size={42} />
+            <p>Generated speech will appear here.</p>
+          </div>
+        )}
+      </section>
+    </div>
+  );
+}
diff --git a/app/frontend/src/services/api.js b/app/frontend/src/services/api.js
index 9c267c1..ce3588e 100644
--- a/app/frontend/src/services/api.js
+++ b/app/frontend/src/services/api.js
@@ -663,6 +663,95 @@ export async function deleteSpeechTranscription(filename) {
   return await readJsonResponse(res, "The local server returned an invalid speech transcription delete response.");
 }
 
+export async function getTtsStatus() {
+  try {
+    const res = await fetch("/api/tts/status");
+    return await readJsonResponse(res, "The local server returned invalid TTS runtime status.");
+  } catch (err) {
+    return { ready: false, running: false, runtimeInstalled: false, error: err.message, settings: {}, voices: [] };
+  }
+}
+
+export async function listTtsModels() {
+  const res = await fetch("/api/tts/models");
+  const data = await readJsonResponse(res, "The local server returned invalid TTS model data.");
+  return data.models || [];
+}
+
+export async function startTts(model, options = {}) {
+  const res = await fetch("/api/tts/start", {
+    method: "POST",
+    headers: { "Content-Type": "application/json" },
+    body: JSON.stringify({
+      model,
+      voice: options.voice,
+      speed: options.speed,
+    }),
+  });
+  return await readJsonResponse(res, "The local server returned an invalid TTS start response.");
+}
+
+export async function stopTts() {
+  const res = await fetch("/api/tts/stop", { method: "POST" });
+  return await readJsonResponse(res, "The local server returned an invalid TTS stop response.");
+}
+
+export async function speakTts(text, options = {}) {
+  const res = await fetch("/api/tts/speak", {
+    method: "POST",
+    headers: { "Content-Type": "application/json" },
+    body: JSON.stringify({
+      text,
+      model: options.model,
+      voice: options.voice,
+      speed: options.speed,
+    }),
+    signal: options.signal,
+  });
+  const data = await readJsonResponse(res, "The local server returned an invalid TTS response.");
+  return data.output;
+}
+
+export async function downloadTtsModel(modelIdOrUrl, filename = null) {
+  const res = await fetch("/api/tts/download-model", {
+    method: "POST",
+    headers: { "Content-Type": "application/json" },
+    body: JSON.stringify({
+      model: modelIdOrUrl,
+      filename,
+    }),
+  });
+  return await readJsonResponse(res, "The local server returned an invalid TTS download response.");
+}
+
+export async function importTtsModel(file, onProgress, signal) {
+  return await uploadModelFileToEndpoint(file, "/api/tts/import-model", onProgress, signal);
+}
+
+export async function deleteTtsModel(filename) {
+  const res = await fetch("/api/tts/delete-model", {
+    method: "POST",
+    headers: { "Content-Type": "application/json" },
+    body: JSON.stringify({ filename }),
+  });
+  return await readJsonResponse(res, "The local server returned an invalid TTS model delete response.");
+}
+
+export async function listTtsOutputs() {
+  const res = await fetch("/api/tts/outputs");
+  const data = await readJsonResponse(res, "The local server returned invalid TTS output history.");
+  return data.outputs || [];
+}
+
+export async function deleteTtsOutput(filename) {
+  const res = await fetch("/api/tts/delete-output", {
+    method: "POST",
+    headers: { "Content-Type": "application/json" },
+    body: JSON.stringify({ filename }),
+  });
+  return await readJsonResponse(res, "The local server returned an invalid TTS output delete response.");
+}
+
 export async function getLlmRecommendations(useCase = "chat", limit = 10) {
   try {
     const res = await fetch(`/api/llm/recommend?useCase=${encodeURIComponent(useCase)}&limit=${limit}`);
diff --git a/linux.sh b/linux.sh
index b58cd9b..79c19ca 100755
--- a/linux.sh
+++ b/linux.sh
@@ -134,6 +134,7 @@ LLM_SYCL_PATH="$APP_DIR/llm-backend/linux/sycl/llama-server"
 LLM_VULKAN_PATH="$APP_DIR/llm-backend/linux/vulkan/llama-server"
 LLM_CPU_PATH="$APP_DIR/llm-backend/linux/cpu/llama-server"
 SPEECH_BACKEND_PATH="$APP_DIR/speech-backend/linux/whisper-cli"
+TTS_RUNTIME_PATH="$APP_DIR/tts-runtime/node_modules/kokoro-js"
 if [[ ! -x "$CPU_BACKEND_PATH" || ! -x "$CPU_SERVER_PATH" ]] && [[ ! -x "$BACKEND_PATH" || ! -x "$VULKAN_SERVER_PATH" ]]; then
   SETUP_REASON="Linux backend binaries are missing or not executable."
 fi
@@ -143,6 +144,9 @@ fi
 if [[ ! -x "$SPEECH_BACKEND_PATH" ]]; then
   SETUP_REASON="Linux whisper.cpp speech backend is missing or not executable."
 fi
+if [[ ! -d "$TTS_RUNTIME_PATH" ]]; then
+  SETUP_REASON="Kokoro text-to-speech runtime is missing."
+fi
 
 if [[ -n "$SETUP_REASON" ]]; then
   echo ""
@@ -221,6 +225,7 @@ echo "   Web UI:     http://localhost:${FRONTEND_PORT}"
 echo "   GPU API:    Auto-selected by the app (starts at 8080)"
 echo "   Text API:   Starts when a GGUF model is loaded (port ${LLM_PORT})"
 echo "   Speech:     Managed locally by the app"
+echo "   TTS:        Managed locally by the app"
 echo ""
 echo "   Press Ctrl+C in this window to stop all services."
 echo "  ============================================================"
diff --git a/mac.sh b/mac.sh
index cc75ef9..1714711 100755
--- a/mac.sh
+++ b/mac.sh
@@ -23,6 +23,7 @@ if [[ "$(uname -m)" == "arm64" ]]; then
 else
   LLM_BACKEND_PATH="$APP_DIR/llm-backend/mac/x64/llama-server"
 fi
+TTS_RUNTIME_PATH="$APP_DIR/tts-runtime/node_modules/kokoro-js"
 PLATFORM_LABEL="macOS"
 
 DIST_INDEX="$APP_DIR/dist/index.html"
@@ -111,6 +112,9 @@ fi
 if [[ ! -x "$LLM_BACKEND_PATH" ]]; then
   SETUP_REASON="No macOS llama.cpp text backend is installed."
 fi
+if [[ ! -d "$TTS_RUNTIME_PATH" ]]; then
+  SETUP_REASON="Kokoro text-to-speech runtime is missing."
+fi
 
 if [[ -n "$SETUP_REASON" ]]; then
   echo ""
@@ -188,7 +192,8 @@ echo "   Running!"
 echo "   Web UI:     http://localhost:${FRONTEND_PORT}"
 echo "   GPU API:    Auto-selected by the app (starts at 8080)"
 echo "   Text API:   Starts when a GGUF model is loaded (port ${LLM_PORT})"
-echo "   Speech:     Available when a whisper.cpp binary is present"
+echo "   Speech:     Managed locally by the app"
+echo "   TTS:        Managed locally by the app"
 echo ""
 echo "   Press Ctrl+C in this window to stop all services."
 echo "  ============================================================"
diff --git a/scripts/reset.ps1 b/scripts/reset.ps1
index 725d08f..ed6a8f1 100644
--- a/scripts/reset.ps1
+++ b/scripts/reset.ps1
@@ -39,6 +39,13 @@ if (Test-Path $speechBackendDir) {
     Remove-Item $speechBackendDir -Recurse -Force -ErrorAction SilentlyContinue
 }
 
+# Delete Kokoro TTS runtime dependencies
+$ttsRuntimeDir = Join-Path $appDir "tts-runtime"
+if (Test-Path $ttsRuntimeDir) {
+    Write-Host "   >> Removing Kokoro TTS runtime dependencies..." -ForegroundColor Cyan
+    Remove-Item $ttsRuntimeDir -Recurse -Force -ErrorAction SilentlyContinue
+}
+
 # Delete dist
 $distDir = Join-Path $appDir "dist"
 if (Test-Path $distDir) {
@@ -64,12 +71,28 @@ if (Test-Path $speechModelsDir) {
     Write-Host "   >> Preserving speech models in app/speech-models." -ForegroundColor Cyan
 }
 
+# Preserve TTS models
+$ttsModelsDir = Join-Path $appDir "tts-models"
+if (Test-Path $ttsModelsDir) {
+    Write-Host "   >> Preserving TTS models in app/tts-models." -ForegroundColor Cyan
+}
+
 # Preserve transcriptions
 $transcriptionsDir = Join-Path $appDir "transcriptions"
 if (Test-Path $transcriptionsDir) {
     Write-Host "   >> Preserving speech transcripts in app/transcriptions." -ForegroundColor Cyan
 }
 
+# Preserve TTS outputs/cache
+$ttsOutputsDir = Join-Path $appDir "tts-outputs"
+if (Test-Path $ttsOutputsDir) {
+    Write-Host "   >> Preserving TTS outputs in app/tts-outputs." -ForegroundColor Cyan
+}
+$ttsCacheDir = Join-Path $appDir "tts-cache"
+if (Test-Path $ttsCacheDir) {
+    Write-Host "   >> Preserving TTS model cache in app/tts-cache." -ForegroundColor Cyan
+}
+
 # Preserve OpenVINO models
 $openVinoModelsDir = Join-Path $appDir "openvino-models"
 if (Test-Path $openVinoModelsDir) {
@@ -101,7 +124,7 @@ if (Test-Path $lockFile) {
 
 Write-Host ""
 Write-Host "  ============================================================" -ForegroundColor Green
-Write-Host "   Reset complete. Image models, text models, speech models, transcripts, OpenVINO models, and generated outputs were preserved." -ForegroundColor Green
+Write-Host "   Reset complete. Image models, text models, speech models, TTS models, transcripts, TTS outputs, OpenVINO models, and generated outputs were preserved." -ForegroundColor Green
 Write-Host "  ============================================================" -ForegroundColor Green
 Write-Host ""
 Read-Host "  Press Enter to close..."
diff --git a/scripts/reset.sh b/scripts/reset.sh
index 3262728..7de4ec6 100755
--- a/scripts/reset.sh
+++ b/scripts/reset.sh
@@ -40,6 +40,12 @@ if [[ -d "$APP_DIR/speech-backend" ]]; then
   rm -rf "$APP_DIR/speech-backend"
 fi
 
+# Delete Kokoro TTS runtime dependencies
+if [[ -d "$APP_DIR/tts-runtime" ]]; then
+  echo "   >> Removing Kokoro TTS runtime dependencies..."
+  rm -rf "$APP_DIR/tts-runtime"
+fi
+
 # Delete dist
 if [[ -d "$APP_DIR/dist" ]]; then
   echo "   >> Removing dist/ build folder..."
@@ -61,11 +67,24 @@ if [[ -d "$APP_DIR/speech-models" ]]; then
   echo "   >> Preserving speech models in app/speech-models."
 fi
 
+# Preserve TTS models
+if [[ -d "$APP_DIR/tts-models" ]]; then
+  echo "   >> Preserving TTS models in app/tts-models."
+fi
+
 # Preserve transcriptions
 if [[ -d "$APP_DIR/transcriptions" ]]; then
   echo "   >> Preserving speech transcripts in app/transcriptions."
 fi
 
+# Preserve TTS outputs/cache
+if [[ -d "$APP_DIR/tts-outputs" ]]; then
+  echo "   >> Preserving TTS outputs in app/tts-outputs."
+fi
+if [[ -d "$APP_DIR/tts-cache" ]]; then
+  echo "   >> Preserving TTS model cache in app/tts-cache."
+fi
+
 # Preserve OpenVINO models
 if [[ -d "$APP_DIR/openvino-models" ]]; then
   echo "   >> Preserving OpenVINO models in app/openvino-models."
@@ -92,7 +111,7 @@ fi
 
 echo ""
 echo "  ============================================================"
-echo "   Reset complete. Image models, text models, speech models, transcripts, OpenVINO models, and generated outputs were preserved."
+echo "   Reset complete. Image models, text models, speech models, TTS models, transcripts, TTS outputs, OpenVINO models, and generated outputs were preserved."
 echo "  ============================================================"
 echo ""
 read -rp "  Press Enter to close..."
diff --git a/scripts/serve.cjs b/scripts/serve.cjs
index 9055b68..7746461 100644
--- a/scripts/serve.cjs
+++ b/scripts/serve.cjs
@@ -31,9 +31,11 @@ const PORT_FRONTEND = readPort(process.env.PORT || process.env.FRONTEND_PORT, 14
 const PREFERRED_BACKEND_PORT = readPort(process.env.BACKEND_PORT || process.env.SD_BACKEND_PORT, 8080);
 const PREFERRED_LLM_PORT = readPort(process.env.LLM_PORT, 10086);
 const PREFERRED_SPEECH_PORT = readPort(process.env.SPEECH_PORT, 10088);
+const PREFERRED_TTS_PORT = readPort(process.env.TTS_PORT, 10089);
 let PORT_BACKEND = PREFERRED_BACKEND_PORT;
 let PORT_LLM = PREFERRED_LLM_PORT;
 let PORT_SPEECH = PREFERRED_SPEECH_PORT;
+let PORT_TTS = PREFERRED_TTS_PORT;
 const MAX_JSON_BODY_BYTES = 64 * 1024 * 1024;
 const SERVER_BUILD = "text-image-v1";
 const ROOT    = path.join(__dirname, "..");
@@ -115,6 +117,20 @@ const TRANSCRIPTIONS = path.join(ROOT, "app", "transcriptions");
 if (!fs.existsSync(TRANSCRIPTIONS)) {
   fs.mkdirSync(TRANSCRIPTIONS, { recursive: true });
 }
+const TTS_MODELS = path.join(ROOT, "app", "tts-models");
+if (!fs.existsSync(TTS_MODELS)) {
+  fs.mkdirSync(TTS_MODELS, { recursive: true });
+}
+const TTS_OUTPUTS = path.join(ROOT, "app", "tts-outputs");
+if (!fs.existsSync(TTS_OUTPUTS)) {
+  fs.mkdirSync(TTS_OUTPUTS, { recursive: true });
+}
+const TTS_CACHE = path.join(ROOT, "app", "tts-cache");
+if (!fs.existsSync(TTS_CACHE)) {
+  fs.mkdirSync(TTS_CACHE, { recursive: true });
+}
+const TTS_RUNTIME = path.join(ROOT, "app", "tts-runtime");
+const TTS_WORKER = path.join(ROOT, "scripts", "tts-kokoro-worker.mjs");
 const SPEECH_BACKEND_PATHS = {
   winCli: path.join(ROOT, "app", "speech-backend", "win", "whisper-cli.exe"),
   winServer: path.join(ROOT, "app", "speech-backend", "win", "whisper-server.exe"),
@@ -135,6 +151,38 @@ const SPEECH_MODEL_CATALOG = [
   { id: "base-q5_1", name: "Whisper Base Multilingual Q5", filename: "ggml-base-q5_1.bin", size: "57 MB", language: "Multilingual", recommended: false },
   { id: "small", name: "Whisper Small Multilingual", filename: "ggml-small.bin", size: "466 MB", language: "Multilingual", recommended: false },
 ];
+const TTS_MODEL_CATALOG = [
+  {
+    id: "kokoro-onnx-q8",
+    name: "Kokoro 82M ONNX Q8",
+    filename: "kokoro-onnx-q8.json",
+    modelId: "onnx-community/Kokoro-82M-v1.0-ONNX",
+    dtype: "q8",
+    size: "Model cache managed by kokoro-js",
+    format: "Kokoro ONNX",
+    recommended: true,
+  },
+  {
+    id: "kokoro-onnx-fp32",
+    name: "Kokoro 82M ONNX FP32",
+    filename: "kokoro-onnx-fp32.json",
+    modelId: "onnx-community/Kokoro-82M-v1.0-ONNX",
+    dtype: "fp32",
+    size: "Model cache managed by kokoro-js",
+    format: "Kokoro ONNX",
+    recommended: false,
+  },
+];
+const TTS_VOICES = [
+  { id: "af_heart", name: "Heart", language: "en-us", gender: "Female", recommended: true },
+  { id: "af_bella", name: "Bella", language: "en-us", gender: "Female", recommended: false },
+  { id: "af_nicole", name: "Nicole", language: "en-us", gender: "Female", recommended: false },
+  { id: "af_sarah", name: "Sarah", language: "en-us", gender: "Female", recommended: false },
+  { id: "am_michael", name: "Michael", language: "en-us", gender: "Male", recommended: false },
+  { id: "am_fenrir", name: "Fenrir", language: "en-us", gender: "Male", recommended: false },
+  { id: "bf_emma", name: "Emma", language: "en-gb", gender: "Female", recommended: false },
+  { id: "bm_george", name: "George", language: "en-gb", gender: "Male", recommended: false },
+];
 const OPENVINO_MODELS = path.join(ROOT, "app", "openvino-models");
 if (!fs.existsSync(OPENVINO_MODELS)) {
   fs.mkdirSync(OPENVINO_MODELS, { recursive: true });
@@ -189,6 +237,24 @@ let speechTranscriptionState = {
   model: "",
   filename: "",
 };
+let ttsReady = false;
+let ttsError = null;
+let ttsOperationQueue = Promise.resolve();
+let ttsSettings = {
+  model: null,
+  voice: "af_heart",
+  speed: 1,
+  dtype: "q8",
+  backendMode: "Kokoro ONNX",
+};
+let ttsGenerationState = {
+  active: false,
+  phase: "",
+  progress: 0,
+  model: "",
+  voice: "",
+  output: "",
+};
 let llmSettings = {
   model: null,
   threads: Math.max(1, Math.min(16, os.cpus().length || 4)),
@@ -1674,6 +1740,300 @@ function resolveSpeechModel(value) {
   return { ...model, path: modelPath };
 }
 
+function getTtsManifestPath(filename) {
+  const manifestPath = path.join(TTS_MODELS, path.basename(String(filename || "")));
+  return pathInside(manifestPath, TTS_MODELS) ? manifestPath : "";
+}
+
+function readTtsManifest(filename) {
+  const manifestPath = getTtsManifestPath(filename);
+  if (!manifestPath || !fs.existsSync(manifestPath)) return null;
+  try {
+    return JSON.parse(fs.readFileSync(manifestPath, "utf8"));
+  } catch (_) {
+    return null;
+  }
+}
+
+function installTtsCatalogModel(modelIdOrFilename) {
+  const raw = String(modelIdOrFilename || "").trim();
+  const catalogModel = TTS_MODEL_CATALOG.find((model) => model.id === raw || model.filename === raw) || TTS_MODEL_CATALOG[0];
+  if (!catalogModel) throw new Error("Unknown TTS model.");
+  const manifestPath = getTtsManifestPath(catalogModel.filename);
+  const manifest = {
+    ...catalogModel,
+    installed: true,
+    createdAt: new Date().toISOString(),
+    notes: "Kokoro model files are cached under app/tts-cache by kokoro-js on first generation.",
+  };
+  fs.writeFileSync(manifestPath, JSON.stringify(manifest, null, 2), "utf8");
+  return manifest;
+}
+
+function getTtsModels() {
+  const catalog = TTS_MODEL_CATALOG.map((model) => {
+    const manifest = readTtsManifest(model.filename);
+    const installed = Boolean(manifest);
+    const manifestPath = getTtsManifestPath(model.filename);
+    const sizeBytes = installed ? getPathSize(manifestPath) : 0;
+    return {
+      ...model,
+      ...(manifest || {}),
+      installed,
+      sizeBytes,
+      localSize: installed ? formatBytes(sizeBytes) : "",
+      url: `kokoro://install/${model.id}`,
+    };
+  });
+
+  const known = new Set(catalog.map((model) => model.filename.toLowerCase()));
+  let custom = [];
+  try {
+    custom = fs.readdirSync(TTS_MODELS)
+      .filter((filename) => filename.toLowerCase().endsWith(".json") && !known.has(filename.toLowerCase()))
+      .map((filename) => {
+        const manifest = readTtsManifest(filename) || {};
+        const sizeBytes = getPathSize(path.join(TTS_MODELS, filename));
+        return {
+          id: manifest.id || filename,
+          name: manifest.name || filename,
+          filename,
+          modelId: manifest.modelId || "onnx-community/Kokoro-82M-v1.0-ONNX",
+          dtype: manifest.dtype || "q8",
+          format: manifest.format || "Kokoro ONNX",
+          size: formatBytes(sizeBytes),
+          localSize: formatBytes(sizeBytes),
+          sizeBytes,
+          installed: true,
+          recommended: false,
+          custom: true,
+        };
+      });
+  } catch (_) {}
+  return [...catalog, ...custom];
+}
+
+function resolveTtsModel(value) {
+  const raw = String(value || "").trim();
+  const model = getTtsModels().find((item) => item.id === raw || item.filename === raw) ||
+    getTtsModels().find((item) => item.installed);
+  if (!model) throw new Error("Download or import a Kokoro TTS model from Model Manager first.");
+  const manifestPath = getTtsManifestPath(model.filename);
+  if (!manifestPath || !fs.existsSync(manifestPath)) {
+    throw new Error(`TTS model is not installed: ${model.filename}`);
+  }
+  return { ...model, path: manifestPath };
+}
+
+function getTtsRuntimeStatus() {
+  const nodeModules = path.join(TTS_RUNTIME, "node_modules");
+  const installed = fs.existsSync(path.join(nodeModules, "kokoro-js"));
+  return {
+    installed: installed && fs.existsSync(TTS_WORKER),
+    worker: TTS_WORKER,
+    runtime: TTS_RUNTIME,
+    cache: TTS_CACHE,
+    backendMode: "Kokoro ONNX",
+  };
+}
+
+function runExclusiveTtsOperation(operation) {
+  const run = ttsOperationQueue.catch(() => {}).then(operation);
+  ttsOperationQueue = run.catch(() => {});
+  return run;
+}
+
+async function startTts(settings = {}) {
+  const runtime = getTtsRuntimeStatus();
+  if (!runtime.installed) {
+    throw new Error("Kokoro TTS runtime is not installed. Run scripts/setup-tts for this platform.");
+  }
+  const model = resolveTtsModel(settings.model || ttsSettings.model);
+  await killBackend();
+  await killOpenVinoWorker();
+  await killLlm();
+  await stopSpeech();
+  PORT_TTS = PREFERRED_TTS_PORT;
+  ttsError = null;
+  ttsReady = true;
+  ttsSettings = {
+    ...ttsSettings,
+    model: model.filename,
+    voice: settings.voice || ttsSettings.voice || "af_heart",
+    speed: Math.max(0.5, Math.min(2, Number(settings.speed) || ttsSettings.speed || 1)),
+    dtype: model.dtype || settings.dtype || ttsSettings.dtype || "q8",
+    backendMode: runtime.backendMode,
+  };
+}
+
+async function stopTts() {
+  ttsReady = false;
+  ttsError = null;
+  ttsGenerationState = { active: false, phase: "", progress: 0, model: "", voice: "", output: "" };
+}
+
+function saveTtsOutput(result) {
+  const createdAt = new Date().toISOString();
+  const stamp = createdAt.replace(/[:.]/g, "-");
+  const base = `tts-${stamp}-${safeOutputName(result.voice || "voice")}`;
+  const wavFilename = `${base}.wav`;
+  const jsonFilename = `${base}.json`;
+  const wavPath = path.join(TTS_OUTPUTS, wavFilename);
+  fs.copyFileSync(result.wavPath, wavPath);
+  const metadata = {
+    text: result.text || "",
+    model: result.model,
+    modelName: result.modelName,
+    voice: result.voice,
+    voiceName: result.voiceName,
+    speed: result.speed,
+    durationMs: result.durationMs,
+    sampleRate: result.sampleRate || 24000,
+    createdAt,
+    audioFile: wavFilename,
+    metadata: jsonFilename,
+    displayName: result.displayName || `${result.voiceName || result.voice} - ${createdAt}`,
+  };
+  fs.writeFileSync(path.join(TTS_OUTPUTS, jsonFilename), JSON.stringify(metadata, null, 2), "utf8");
+  return {
+    ...metadata,
+    filename: jsonFilename,
+    url: `/tts-outputs/${encodeURIComponent(wavFilename)}`,
+  };
+}
+
+function listTtsOutputs() {
+  try {
+    return fs.readdirSync(TTS_OUTPUTS)
+      .filter((file) => file.toLowerCase().endsWith(".json"))
+      .map((file) => {
+        const filePath = path.join(TTS_OUTPUTS, file);
+        try {
+          const metadata = JSON.parse(fs.readFileSync(filePath, "utf8"));
+          const stat = fs.statSync(filePath);
+          return {
+            ...metadata,
+            filename: file,
+            displayName: metadata.displayName || metadata.text?.slice(0, 48) || metadata.audioFile || file,
+            sizeBytes: stat.size,
+            size: formatBytes(stat.size),
+            modifiedAt: stat.mtime.toISOString(),
+            createdAt: metadata.createdAt || stat.mtime.toISOString(),
+            url: metadata.audioFile ? `/tts-outputs/${encodeURIComponent(metadata.audioFile)}` : "",
+          };
+        } catch (_) {
+          return null;
+        }
+      })
+      .filter(Boolean)
+      .sort((a, b) => String(b.createdAt || "").localeCompare(String(a.createdAt || "")));
+  } catch (_) {
+    return [];
+  }
+}
+
+function synthesizeTts(text, options = {}) {
+  return new Promise((resolve, reject) => {
+    const cleanedText = String(text || "").trim();
+    if (!cleanedText) {
+      reject(new Error("Enter text to synthesize."));
+      return;
+    }
+    if (cleanedText.length > 5000) {
+      reject(new Error("TTS text is too long. Limit is 5000 characters for V1."));
+      return;
+    }
+    const runtime = getTtsRuntimeStatus();
+    if (!runtime.installed) {
+      reject(new Error("Kokoro TTS runtime is not installed."));
+      return;
+    }
+    const model = resolveTtsModel(options.model || ttsSettings.model);
+    const voice = String(options.voice || ttsSettings.voice || "af_heart");
+    const voiceInfo = TTS_VOICES.find((item) => item.id === voice) || { id: voice, name: voice };
+    const speed = Math.max(0.5, Math.min(2, Number(options.speed) || ttsSettings.speed || 1));
+    const stamp = new Date().toISOString().replace(/[:.]/g, "-");
+    const tempWav = path.join(TTS_OUTPUTS, `.tts-${stamp}.wav`);
+    const payload = {
+      text: cleanedText,
+      modelId: model.modelId || "onnx-community/Kokoro-82M-v1.0-ONNX",
+      dtype: model.dtype || "q8",
+      voice,
+      speed,
+      output: tempWav,
+      cacheDir: TTS_CACHE,
+    };
+
+    ttsGenerationState = {
+      active: true,
+      phase: "Generating speech...",
+      progress: -1,
+      model: model.filename,
+      voice,
+      output: "",
+    };
+
+    const startedAt = Date.now();
+    const proc = spawn(process.execPath, [TTS_WORKER], {
+      cwd: ROOT,
+      stdio: "pipe",
+      windowsHide: true,
+      env: {
+        ...process.env,
+        NODE_PATH: path.join(TTS_RUNTIME, "node_modules"),
+        TTS_RUNTIME,
+      },
+    });
+    let stdout = "";
+    let stderr = "";
+    proc.stdout.on("data", (data) => { stdout += data.toString(); });
+    proc.stderr.on("data", (data) => {
+      stderr += data.toString();
+      process.stderr.write("  [tts] " + data.toString());
+    });
+    proc.on("error", (err) => {
+      ttsGenerationState = { active: false, phase: "Failed", progress: 0, model: model.filename, voice, output: "" };
+      reject(err);
+    });
+    proc.on("exit", (code) => {
+      ttsGenerationState = { active: false, phase: code === 0 ? "Complete" : "Failed", progress: code === 0 ? 100 : 0, model: model.filename, voice, output: "" };
+      if (code !== 0) {
+        try { fs.unlinkSync(tempWav); } catch (_) {}
+        const message = (stderr || stdout || `Kokoro TTS exited with code ${code}`).trim().slice(-1400);
+        ttsError = message;
+        reject(new Error(message));
+        return;
+      }
+      try {
+        const workerResult = JSON.parse(stdout.trim() || "{}");
+        if (!workerResult.ok || !fs.existsSync(tempWav)) {
+          throw new Error(workerResult.error || "Kokoro TTS did not produce a WAV file.");
+        }
+        const saved = saveTtsOutput({
+          text: cleanedText,
+          model: model.filename,
+          modelName: model.name,
+          voice,
+          voiceName: voiceInfo.name,
+          speed,
+          durationMs: Date.now() - startedAt,
+          sampleRate: workerResult.sampleRate || 24000,
+          wavPath: tempWav,
+          displayName: `${voiceInfo.name || voice} - ${cleanedText.slice(0, 40)}`,
+        });
+        try { fs.unlinkSync(tempWav); } catch (_) {}
+        ttsError = null;
+        resolve(saved);
+      } catch (err) {
+        try { fs.unlinkSync(tempWav); } catch (_) {}
+        ttsError = err.message || String(err);
+        reject(err);
+      }
+    });
+    proc.stdin.end(JSON.stringify(payload));
+  });
+}
+
 function runExclusiveSpeechOperation(operation) {
   const run = speechOperationQueue.catch(() => {}).then(operation);
   speechOperationQueue = run.catch(() => {});
@@ -2091,13 +2451,25 @@ async function getHealth() {
       totalBytes: getPathSize(OUTPUTS),
       totalSize: formatBytes(getPathSize(OUTPUTS)),
     },
+    tts: {
+      models: fs.existsSync(TTS_MODELS) ? fs.readdirSync(TTS_MODELS).filter((file) => file.toLowerCase().endsWith(".json")).length : 0,
+      outputs: fs.existsSync(TTS_OUTPUTS) ? fs.readdirSync(TTS_OUTPUTS).filter((file) => file.toLowerCase().endsWith(".json")).length : 0,
+      runtimeInstalled: getTtsRuntimeStatus().installed,
+      totalBytes: getPathSize(TTS_MODELS) + getPathSize(TTS_OUTPUTS) + getPathSize(TTS_CACHE),
+      totalSize: formatBytes(getPathSize(TTS_MODELS) + getPathSize(TTS_OUTPUTS) + getPathSize(TTS_CACHE)),
+    },
     issues,
   };
 }
 
 function addCleanupCandidate(candidates, id, targetPath, reason, options = {}) {
   if (!fs.existsSync(targetPath)) return;
-  if (!options.allowUserData && (pathInside(targetPath, MODELS) || pathInside(targetPath, OUTPUTS))) return;
+  if (!options.allowUserData && (
+    pathInside(targetPath, MODELS) ||
+    pathInside(targetPath, OUTPUTS) ||
+    pathInside(targetPath, TTS_MODELS) ||
+    pathInside(targetPath, TTS_OUTPUTS)
+  )) return;
   const sizeBytes = getPathSize(targetPath);
   candidates.push({
     id,
@@ -4072,7 +4444,13 @@ function cancelModelDownload() {
       try { fs.unlinkSync(activeDownload.tempPath || activeDownload.destPath); } catch (_) {}
     }
   } else if (filename) {
-    const targetDir = downloadState.kind === "text" ? LLM_MODELS : downloadState.kind === "speech" ? SPEECH_MODELS : MODELS;
+    const targetDir = downloadState.kind === "text"
+      ? LLM_MODELS
+      : downloadState.kind === "speech"
+        ? SPEECH_MODELS
+        : downloadState.kind === "tts"
+          ? TTS_MODELS
+          : MODELS;
     try { fs.unlinkSync(`${path.join(targetDir, filename)}.part`); } catch (_) {}
   }
   activeDownload = null;
@@ -4397,9 +4775,17 @@ function streamModelUpload(req, filename, targetDir = MODELS, mode = "image") {
       ? lowerName.endsWith(".gguf")
       : mode === "speech"
         ? lowerName.endsWith(".bin")
+        : mode === "tts"
+          ? lowerName.endsWith(".json")
         : isModelFile(lowerName);
     if (!safeFilename || !valid) {
-      reject(new Error(mode === "text" ? "Filename must end with .gguf" : mode === "speech" ? "Filename must end with .bin" : "Filename must end with .gguf, .safetensors, or .ckpt"));
+      reject(new Error(mode === "text"
+        ? "Filename must end with .gguf"
+        : mode === "speech"
+          ? "Filename must end with .bin"
+          : mode === "tts"
+            ? "Filename must end with .json"
+            : "Filename must end with .gguf, .safetensors, or .ckpt"));
       return;
     }
 
@@ -4771,6 +5157,136 @@ const server = http.createServer(async (req, res) => {
     }
   }
 
+  if (req.url === "/api/tts/status" && req.method === "GET") {
+    const runtime = getTtsRuntimeStatus();
+    return json(res, 200, {
+      ok: true,
+      ready: ttsReady,
+      running: ttsReady,
+      port: PORT_TTS,
+      preferredPort: PREFERRED_TTS_PORT,
+      runtimeInstalled: runtime.installed,
+      runtimePath: runtime.runtime,
+      workerPath: runtime.worker,
+      cachePath: runtime.cache,
+      backendMode: runtime.backendMode,
+      error: ttsError,
+      settings: ttsSettings,
+      generation: ttsGenerationState,
+      voices: TTS_VOICES,
+    });
+  }
+
+  if (req.url === "/api/tts/models" && req.method === "GET") {
+    return json(res, 200, { ok: true, models: getTtsModels(), voices: TTS_VOICES });
+  }
+
+  if (req.url === "/api/tts/start" && req.method === "POST") {
+    const body = await readJsonBody(req, res);
+    if (!body) return;
+    try {
+      await runExclusiveTtsOperation(() => startTts(body));
+      return json(res, 200, { ok: true, ready: ttsReady, settings: ttsSettings, port: PORT_TTS });
+    } catch (err) {
+      ttsError = err.message || String(err);
+      return json(res, 500, { ok: false, error: ttsError });
+    }
+  }
+
+  if (req.url === "/api/tts/stop" && req.method === "POST") {
+    await stopTts();
+    return json(res, 200, { ok: true });
+  }
+
+  if (req.url === "/api/tts/speak" && req.method === "POST") {
+    const body = await readJsonBody(req, res, 512 * 1024);
+    if (!body) return;
+    try {
+      if (!ttsReady || (body.model && ttsSettings.model !== body.model)) {
+        await runExclusiveTtsOperation(() => startTts(body));
+      }
+      const output = await runExclusiveTtsOperation(() => synthesizeTts(body.text, body));
+      return json(res, 200, { ok: true, output });
+    } catch (err) {
+      ttsError = err.message || String(err);
+      return json(res, 500, { ok: false, error: ttsError });
+    }
+  }
+
+  if (req.url === "/api/tts/download-model" && req.method === "POST") {
+    const body = await readJsonBody(req, res);
+    if (!body) return;
+    try {
+      const model = installTtsCatalogModel(body.modelId || body.model_id || body.model || body.filename);
+      return json(res, 200, { ok: true, message: "TTS model manifest installed", filename: model.filename, model });
+    } catch (err) {
+      return json(res, 500, { ok: false, error: err.message || String(err) });
+    }
+  }
+
+  if (req.url.startsWith("/api/tts/import-model") && req.method === "POST") {
+    try {
+      const parsed = new URL(req.url, `http://${req.headers.host || "localhost"}`);
+      const result = await streamModelUpload(req, parsed.searchParams.get("filename"), TTS_MODELS, "tts");
+      return json(res, 200, { ok: true, message: `Imported ${result.filename}`, model: result });
+    } catch (err) {
+      console.error("  [api] Failed to import TTS model:", err);
+      return json(res, 500, { ok: false, error: err.message });
+    }
+  }
+
+  if (req.url === "/api/tts/delete-model" && req.method === "POST") {
+    const body = await readJsonBody(req, res);
+    if (!body) return;
+    const filename = path.basename(String(body.filename || ""));
+    const modelPath = path.join(TTS_MODELS, filename);
+    if (!filename || !filename.endsWith(".json") || !pathInside(modelPath, TTS_MODELS)) {
+      return json(res, 400, { ok: false, error: "Invalid filename" });
+    }
+    if (ttsSettings.model === filename) await stopTts();
+    try {
+      fs.unlinkSync(modelPath);
+      return json(res, 200, { ok: true });
+    } catch (err) {
+      return json(res, err.code === "ENOENT" ? 404 : 500, { ok: false, error: err.code === "ENOENT" ? "TTS model not found" : err.message });
+    }
+  }
+
+  if (req.url === "/api/tts/outputs" && req.method === "GET") {
+    return json(res, 200, { ok: true, outputs: listTtsOutputs() });
+  }
+
+  if (req.url === "/api/tts/delete-output" && req.method === "POST") {
+    const body = await readJsonBody(req, res);
+    if (!body) return;
+    const filename = path.basename(String(body.filename || ""));
+    if (!filename || !filename.endsWith(".json")) {
+      return json(res, 400, { ok: false, error: "Invalid filename" });
+    }
+    const jsonPath = path.join(TTS_OUTPUTS, filename);
+    if (!pathInside(jsonPath, TTS_OUTPUTS)) {
+      return json(res, 400, { ok: false, error: "Invalid path" });
+    }
+    try {
+      if (!fs.existsSync(jsonPath)) {
+        return json(res, 404, { ok: false, error: "TTS output metadata not found" });
+      }
+      const metadata = JSON.parse(fs.readFileSync(jsonPath, "utf8"));
+      const audioFile = metadata.audioFile ? path.basename(String(metadata.audioFile)) : "";
+      fs.unlinkSync(jsonPath);
+      if (audioFile) {
+        const audioPath = path.join(TTS_OUTPUTS, audioFile);
+        if (pathInside(audioPath, TTS_OUTPUTS) && fs.existsSync(audioPath)) {
+          fs.unlinkSync(audioPath);
+        }
+      }
+      return json(res, 200, { ok: true });
+    } catch (err) {
+      console.error("  [api] Failed to delete TTS output:", err);
+      return json(res, 500, { ok: false, error: err.message });
+    }
+  }
+
 function isOomError(msg) {
   if (!msg) return false;
   const lower = String(msg).toLowerCase();
@@ -5307,6 +5823,21 @@ async function getLlmfitRecommendations(useCase = "chat", limit = 10) {
     return;
   }
 
+  if (req.url.startsWith("/tts-outputs/") && req.method === "GET") {
+    const filename = path.basename(decodeURIComponent(req.url.replace(/^\/tts-outputs\//, "").split("?")[0] || ""));
+    const filePath = path.join(TTS_OUTPUTS, filename);
+    if (!filename || !pathInside(filePath, TTS_OUTPUTS) || !fs.existsSync(filePath)) {
+      return json(res, 404, { ok: false, error: "TTS output not found" });
+    }
+    const ext = path.extname(filePath).toLowerCase();
+    fs.readFile(filePath, (err, data) => {
+      if (err) return json(res, 500, { ok: false, error: err.message });
+      res.writeHead(200, { "Content-Type": MIME[ext] || "application/octet-stream", "Access-Control-Allow-Origin": "*" });
+      res.end(data);
+    });
+    return;
+  }
+
   // POST /api/save-output
   if (req.url === "/api/save-output" && req.method === "POST") {
     try {
@@ -5436,6 +5967,7 @@ server.listen(PORT_FRONTEND, "0.0.0.0", () => {
   console.log("   Image API: http://127.0.0.1:" + PORT_BACKEND);
   console.log("   Text API : starts on http://127.0.0.1:" + PREFERRED_LLM_PORT);
   console.log("   Speech   : managed locally, API on frontend port");
+  console.log("   TTS      : Kokoro ONNX managed locally, API on frontend port");
   console.log("  ============================================================");
   console.log("");
 
@@ -5444,5 +5976,5 @@ server.listen(PORT_FRONTEND, "0.0.0.0", () => {
 });
 
 // Graceful shutdown
-process.on("SIGINT",  async () => { await killBackend(); await killOpenVinoWorker(); await killLlm(); await stopSpeech(); process.exit(0); });
-process.on("SIGTERM", async () => { await killBackend(); await killOpenVinoWorker(); await killLlm(); await stopSpeech(); process.exit(0); });
+process.on("SIGINT",  async () => { await killBackend(); await killOpenVinoWorker(); await killLlm(); await stopSpeech(); await stopTts(); process.exit(0); });
+process.on("SIGTERM", async () => { await killBackend(); await killOpenVinoWorker(); await killLlm(); await stopSpeech(); await stopTts(); process.exit(0); });
diff --git a/scripts/setup-tts.ps1 b/scripts/setup-tts.ps1
new file mode 100644
index 0000000..2bd6223
--- /dev/null
+++ b/scripts/setup-tts.ps1
@@ -0,0 +1,52 @@
+# Local AI Studio - Kokoro TTS setup for Windows
+
+$scriptDir = Split-Path -Parent $MyInvocation.MyCommand.Path
+$rootDir = Split-Path -Parent $scriptDir
+$appDir = Join-Path $rootDir "app"
+$toolsDir = Join-Path $appDir "tools"
+$nodeDir = Join-Path $toolsDir "node-win"
+$nodeExe = Join-Path $nodeDir "node.exe"
+$npmCmd = Join-Path $nodeDir "npm.cmd"
+$runtimeDir = Join-Path $appDir "tts-runtime"
+$modelsDir = Join-Path $appDir "tts-models"
+$outputsDir = Join-Path $appDir "tts-outputs"
+$cacheDir = Join-Path $appDir "tts-cache"
+
+function Print-OK { param([string]$m); Write-Host "   OK  $m" -ForegroundColor Green }
+function Print-Info { param([string]$m); Write-Host "   >>  $m" -ForegroundColor Cyan }
+function Print-Fail { param([string]$m); Write-Host "   XX  $m" -ForegroundColor Red }
+
+Write-Host ""
+Write-Host "  ============================================================" -ForegroundColor Cyan
+Write-Host "   Setting up Kokoro ONNX Text-to-Speech runtime" -ForegroundColor Cyan
+Write-Host "  ============================================================" -ForegroundColor Cyan
+Write-Host ""
+
+New-Item -ItemType Directory -Force -Path $runtimeDir, $modelsDir, $outputsDir, $cacheDir | Out-Null
+
+if (-not (Test-Path $nodeExe) -or -not (Test-Path $npmCmd)) {
+    Print-Fail "Portable Node.js is missing. Run scripts/setup.ps1 first."
+    exit 1
+}
+
+$packageJson = Join-Path $runtimeDir "package.json"
+if (-not (Test-Path $packageJson)) {
+    '{"private":true,"type":"module","dependencies":{"kokoro-js":"^1.2.1"}}' | Out-File -FilePath $packageJson -Encoding utf8
+}
+
+Push-Location $runtimeDir
+$oldPath = $env:PATH
+try {
+    $env:PATH = "$nodeDir;$env:PATH"
+    Print-Info "Installing kokoro-js into app/tts-runtime..."
+    & $npmCmd install --prefer-offline 2>&1
+    if ($LASTEXITCODE -ne 0) {
+        Print-Fail "kokoro-js install failed."
+        exit 1
+    }
+} finally {
+    $env:PATH = $oldPath
+    Pop-Location
+}
+
+Print-OK "Kokoro TTS runtime is ready."
diff --git a/scripts/setup-tts.sh b/scripts/setup-tts.sh
new file mode 100644
index 0000000..46f11c3
--- /dev/null
+++ b/scripts/setup-tts.sh
@@ -0,0 +1,55 @@
+#!/usr/bin/env bash
+#
+# Local AI Studio - Kokoro TTS setup for Linux/macOS
+#
+
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+ROOT_DIR="$(dirname "$SCRIPT_DIR")"
+APP_DIR="$ROOT_DIR/app"
+TOOLS_DIR="$APP_DIR/tools"
+PLATFORM="$(uname -s)"
+
+if [[ "$PLATFORM" == "Darwin" ]]; then
+  NODE_DIR="$TOOLS_DIR/node-mac"
+else
+  NODE_DIR="$TOOLS_DIR/node-linux"
+fi
+
+NODE_BIN="$NODE_DIR/bin/node"
+NPM_BIN="$NODE_DIR/bin/npm"
+RUNTIME_DIR="$APP_DIR/tts-runtime"
+MODELS_DIR="$APP_DIR/tts-models"
+OUTPUTS_DIR="$APP_DIR/tts-outputs"
+CACHE_DIR="$APP_DIR/tts-cache"
+
+print_ok() { echo "   OK   $1"; }
+print_info() { echo "   >>   $1"; }
+print_fail() { echo "   XX   $1"; }
+
+echo ""
+echo "  ============================================================"
+echo "   Setting up Kokoro ONNX Text-to-Speech runtime"
+echo "  ============================================================"
+echo ""
+
+mkdir -p "$RUNTIME_DIR" "$MODELS_DIR" "$OUTPUTS_DIR" "$CACHE_DIR"
+
+if [[ ! -x "$NODE_BIN" || ! -x "$NPM_BIN" ]]; then
+  print_fail "Portable Node.js is missing. Run scripts/setup.sh first."
+  exit 1
+fi
+
+if [[ ! -f "$RUNTIME_DIR/package.json" ]]; then
+  cat > "$RUNTIME_DIR/package.json" <<'JSON'
+{"private":true,"type":"module","dependencies":{"kokoro-js":"^1.2.1"}}
+JSON
+fi
+
+cd "$RUNTIME_DIR"
+export PATH="$NODE_DIR/bin:$PATH"
+print_info "Installing kokoro-js into app/tts-runtime..."
+"$NPM_BIN" install --prefer-offline
+
+print_ok "Kokoro TTS runtime is ready."
diff --git a/scripts/setup.ps1 b/scripts/setup.ps1
index 5eacced..850581f 100644
--- a/scripts/setup.ps1
+++ b/scripts/setup.ps1
@@ -158,7 +158,7 @@ function Expand-WithProgress {
 # ══════════════════════════════════════════════════════════════════════════════
 Print-Header
 
-$steps = 6
+$steps = 7
 
 # ── Step 1: Portable Node.js ──────────────────────────────────────────────────
 Print-Step 1 $steps "Setting up portable Node.js (app/tools/node-win/)"
@@ -434,7 +434,14 @@ if (-not $?) {
     Read-Host; exit 1
 }
 
-Print-Step 5 $steps "Installing frontend dependencies (app/frontend/)"
+Print-Step 5 $steps "Setting up Kokoro ONNX text-to-speech runtime"
+& (Join-Path $scriptDir "setup-tts.ps1")
+if (-not $?) {
+    Print-Fail "Kokoro TTS setup failed."
+    Read-Host; exit 1
+}
+
+Print-Step 6 $steps "Installing frontend dependencies (app/frontend/)"
 Write-Host ""
 
 if (-not (Test-Path $npmCmd)) {
@@ -488,7 +495,7 @@ try {
     Print-OK "Dependencies installed!"
 
     # ── Step 4: Build frontend ────────────────────────────────────────────────
-    Print-Step 6 $steps "Building frontend -> app/dist/"
+    Print-Step 7 $steps "Building frontend -> app/dist/"
     Write-Host ""
 
     & $npmCmd run build 2>&1
diff --git a/scripts/setup.sh b/scripts/setup.sh
index 54411c2..e02f5b5 100755
--- a/scripts/setup.sh
+++ b/scripts/setup.sh
@@ -258,7 +258,7 @@ if [[ "$PLATFORM" == "Linux" ]]; then
   check_glibc
 fi
 
-TOTAL_STEPS=6
+TOTAL_STEPS=7
 
 # ── Step 1: Portable Node.js ────────────────────────────────────────────────
 print_step 1 $TOTAL_STEPS "Setting up portable Node.js ($NODE_DIR/)"
@@ -547,7 +547,10 @@ bash "$SCRIPT_DIR/setup-llama.sh"
 print_step 4 $TOTAL_STEPS "Setting up whisper.cpp speech backend"
 bash "$SCRIPT_DIR/setup-whisper.sh"
 
-print_step 5 $TOTAL_STEPS "Installing frontend dependencies (app/frontend/)"
+print_step 5 $TOTAL_STEPS "Setting up Kokoro ONNX text-to-speech runtime"
+bash "$SCRIPT_DIR/setup-tts.sh"
+
+print_step 6 $TOTAL_STEPS "Installing frontend dependencies (app/frontend/)"
 
 if [[ ! -x "$NPM_BIN" ]]; then
   print_fail "Portable npm was not found at $NPM_BIN"
@@ -645,7 +648,7 @@ else
 fi
 
 # ── Step 4: Build frontend ──────────────────────────────────────────────────
-print_step 6 $TOTAL_STEPS "Building frontend -> app/dist/"
+print_step 7 $TOTAL_STEPS "Building frontend -> app/dist/"
 
 if [ "$USE_SYMLINKS" = false ]; then
   # If symlinks are disabled, run vite directly using the local node executable.
diff --git a/scripts/tts-kokoro-worker.mjs b/scripts/tts-kokoro-worker.mjs
new file mode 100644
index 0000000..359df16
--- /dev/null
+++ b/scripts/tts-kokoro-worker.mjs
@@ -0,0 +1,59 @@
+import { createRequire } from "node:module";
+import fs from "node:fs/promises";
+import path from "node:path";
+
+const require = createRequire(import.meta.url);
+
+function readStdin() {
+  return new Promise((resolve, reject) => {
+    let data = "";
+    process.stdin.setEncoding("utf8");
+    process.stdin.on("data", (chunk) => { data += chunk; });
+    process.stdin.on("end", () => resolve(data));
+    process.stdin.on("error", reject);
+  });
+}
+
+function resolveRuntimePackage(name) {
+  const runtime = process.env.TTS_RUNTIME || "";
+  if (runtime) {
+    const pkg = path.join(runtime, "node_modules", name, "dist", "kokoro.js");
+    return pathToFileUrl(pkg);
+  }
+  return name;
+}
+
+function pathToFileUrl(filePath) {
+  const resolved = path.resolve(filePath).replace(/\\/g, "/");
+  return `file:///${resolved.replace(/^\/+/, "")}`;
+}
+
+async function main() {
+  const payload = JSON.parse(await readStdin());
+  process.env.TRANSFORMERS_CACHE = payload.cacheDir || process.env.TRANSFORMERS_CACHE || "";
+  process.env.HF_HOME = payload.cacheDir || process.env.HF_HOME || "";
+
+  const kokoro = await import(resolveRuntimePackage("kokoro-js"));
+  const { KokoroTTS } = kokoro;
+
+  const tts = await KokoroTTS.from_pretrained(payload.modelId, {
+    dtype: payload.dtype || "q8",
+    device: "cpu",
+  });
+  const audio = await tts.generate(payload.text, {
+    voice: payload.voice || "af_heart",
+    speed: Number(payload.speed) || 1,
+  });
+  await fs.mkdir(path.dirname(payload.output), { recursive: true });
+  audio.save(payload.output);
+  process.stdout.write(JSON.stringify({
+    ok: true,
+    output: payload.output,
+    sampleRate: audio.sampling_rate || audio.sampleRate || 24000,
+  }));
+}
+
+main().catch((err) => {
+  process.stderr.write(err?.stack || err?.message || String(err));
+  process.exit(1);
+});
diff --git a/windows.bat b/windows.bat
index 0467ed5..e7f7712 100644
--- a/windows.bat
+++ b/windows.bat
@@ -15,6 +15,7 @@ set LLM_VULKAN_BACKEND=%APP%\llm-backend\win\vulkan\llama-server.exe
 set LLM_SYCL_BACKEND=%APP%\llm-backend\win\sycl\llama-server.exe
 set LLM_CPU_BACKEND=%APP%\llm-backend\win\cpu\llama-server.exe
 set SPEECH_BACKEND=%APP%\speech-backend\win\whisper-cli.exe
+set TTS_RUNTIME=%APP%\tts-runtime\node_modules\kokoro-js
 set SERVE=%~dp0scripts\serve.cjs
 if "%FRONTEND_PORT%"=="" set FRONTEND_PORT=1420
 if "%LLM_PORT%"=="" set LLM_PORT=10086
@@ -43,6 +44,10 @@ if not exist "%SPEECH_BACKEND%" (
     set SETUP_REASON=whisper.cpp speech backend is missing.
     goto :run_setup
 )
+if not exist "%TTS_RUNTIME%" (
+    set SETUP_REASON=Kokoro text-to-speech runtime is missing.
+    goto :run_setup
+)
 if exist "%CUDA_BACKEND%" goto :launch
 if exist "%VULKAN_BACKEND%" goto :launch
 set SETUP_REASON=No backend binary is installed.
@@ -107,6 +112,7 @@ echo   Web UI:     http://localhost:%FRONTEND_PORT%
 echo   GPU API:    Auto-selected by the app (starts at 8080)
 echo   Text API:   Starts when a GGUF model is loaded (port %LLM_PORT%)
 echo   Speech:     Managed locally by the app
+echo   TTS:        Managed locally by the app
 echo.
 echo   Press Ctrl+C in this window to stop all services.
 echo  ============================================================