ui: show generation statistics by default in chat settings (#26624)

* chore: Upgrade Storybook

* chore: Bump package-lock

* chore: bump vitest to 4.1.10

* ui: bump fast-uri to 3.1.5

* ui: bump ip-address to 10.4.0

* ui: bump js-yaml to 4.3.1

* ui: bump immutable to 5.1.9

* ui: bump postcss to 8.5.25

* ui: bump brace-expansion to safe versions

* ui: bump sharp to 0.35.3 via override

* ui: bump body-parser to 2.3.0

* ui: bump vite to 7.3.6 and esbuild to 0.28.1

Assisted-by: Claude Sonnet

* ui: bump hono to 4.13.0

* ui: bump dompurify to 3.4.13

* ui: bump @sveltejs/kit to 2.70.2

* ui: bump @modelcontextprotocol/sdk to 1.30.0

* ui: bump valibot to 1.4.2 via override

* chore: Remove legacy setup file

* refactor: Nits cleanup

.github/workflows/build-apple.yml

@@ -60,7 +60,6 @@ jobs:
             -DCMAKE_BUILD_RPATH="@loader_path" \
             -DLLAMA_FATAL_WARNINGS=ON \
             -DLLAMA_BUILD_BORINGSSL=ON \
-            -DGGML_METAL_USE_BF16=ON \
             -DGGML_METAL_EMBED_LIBRARY=OFF \
             -DGGML_METAL_SHADER_DEBUG=ON \
             -DGGML_RPC=ON \
@@ -127,7 +126,6 @@ jobs:
         run: |
           sysctl -a
           cmake -B build -G Xcode \
-            -DGGML_METAL_USE_BF16=ON \
             -DGGML_METAL_EMBED_LIBRARY=ON \
             -DLLAMA_OPENSSL=OFF \
             -DLLAMA_BUILD_APP=OFF \
@@ -178,7 +176,6 @@ jobs:
         run: |
           sysctl -a
           cmake -B build -G Xcode \
-            -DGGML_METAL_USE_BF16=ON \
             -DGGML_METAL_EMBED_LIBRARY=ON \
             -DLLAMA_BUILD_COMMON=OFF \
             -DLLAMA_BUILD_APP=OFF \
@@ -212,7 +209,6 @@ jobs:
         run: |
           sysctl -a
           cmake -B build -G Xcode \
-            -DGGML_METAL_USE_BF16=ON \
             -DGGML_METAL_EMBED_LIBRARY=ON \
             -DLLAMA_BUILD_COMMON=OFF \
             -DLLAMA_BUILD_APP=OFF \
@@ -257,7 +253,6 @@ jobs:
         run: |
           sysctl -a
           cmake -B build -G Xcode \
-            -DGGML_METAL_USE_BF16=ON \
             -DGGML_METAL_EMBED_LIBRARY=ON \
             -DLLAMA_OPENSSL=OFF \
             -DLLAMA_BUILD_APP=OFF \

.github/workflows/release.yml

@@ -93,13 +93,13 @@ jobs:
           - build: 'arm64'
             arch: 'arm64'
             os: macos-26
-            defines: "-DGGML_METAL_USE_BF16=ON -DGGML_METAL_EMBED_LIBRARY=ON -DCMAKE_OSX_DEPLOYMENT_TARGET=13.3"
+            defines: "-DGGML_METAL_EMBED_LIBRARY=ON -DCMAKE_OSX_DEPLOYMENT_TARGET=13.3"
           # TODO: this build is disabled to save Github Actions resources (https://github.com/ggml-org/llama.cpp/pull/23780)
           #       in order to enable it again, we have to provision dedicated runners  to run it
           #- build: 'arm64-kleidiai'
           #  arch: 'arm64'
           #  os: macos-14
-          #  defines: "-DGGML_METAL_USE_BF16=ON -DGGML_METAL_EMBED_LIBRARY=ON -DCMAKE_OSX_DEPLOYMENT_TARGET=13.3 -DGGML_CPU_KLEIDIAI=ON"
+          #  defines: "-DGGML_METAL_EMBED_LIBRARY=ON -DCMAKE_OSX_DEPLOYMENT_TARGET=13.3 -DGGML_CPU_KLEIDIAI=ON"
           - build: 'x64'
             arch: 'x64'
             os: macos-15-intel
@@ -1402,7 +1402,6 @@ jobs:
         run: |
           sysctl -a
           cmake -B build -G Xcode \
-            -DGGML_METAL_USE_BF16=ON \
             -DGGML_METAL_EMBED_LIBRARY=ON \
             -DLLAMA_OPENSSL=OFF \
             -DLLAMA_BUILD_APP=OFF \

build-xcframework.sh

@@ -17,7 +17,6 @@ LLAMA_BUILD_MTMD=ON
 GGML_METAL=ON
 GGML_METAL_EMBED_LIBRARY=ON
 GGML_BLAS_DEFAULT=ON
-GGML_METAL_USE_BF16=ON
 GGML_OPENMP=OFF
 
 COMMON_C_FLAGS="-Wno-macro-redefined -Wno-shorten-64-to-32 -Wno-unused-command-line-argument -g"
@@ -44,7 +43,6 @@ COMMON_CMAKE_ARGS=(
     -DGGML_METAL_EMBED_LIBRARY=${GGML_METAL_EMBED_LIBRARY}
     -DGGML_BLAS_DEFAULT=${GGML_BLAS_DEFAULT}
     -DGGML_METAL=${GGML_METAL}
-    -DGGML_METAL_USE_BF16=${GGML_METAL_USE_BF16}
     -DGGML_NATIVE=OFF
     -DGGML_OPENMP=${GGML_OPENMP}
 )

scripts/ui-assets.cmake

@@ -123,15 +123,15 @@ function(npm_build out_var)
     endif()
 
     if(need_install)
-        message(STATUS "UI: running npm install")
+        message(STATUS "UI: running npm ci")
         execute_process(
-            COMMAND ${NPM_EXECUTABLE} install
+            COMMAND ${NPM_EXECUTABLE} ci
             WORKING_DIRECTORY "${WORK_DIR}"
             RESULT_VARIABLE rc
             ERROR_VARIABLE  err
         )
         if(NOT rc EQUAL 0)
-            message(STATUS "UI: npm install failed (${rc})")
+            message(STATUS "UI: npm ci failed (${rc})")
             message(STATUS "  stderr: ${err}")
             return()
         endif()

tools/mtmd/mtmd-audio.cpp

@@ -556,10 +556,8 @@ bool mtmd_audio_preprocessor_whisper::preprocess(const float *                 s
     }
 
     std::vector<float> smpl;
-    // if input is too short, pad with zeros
-    // this is to avoid potential issues with stage1/2 padding in log_mel_spectrogram
-    // TODO: maybe handle this better
-    size_t min_samples = (size_t) hparams.audio_sample_rate * (hparams.audio_chunk_len + 1);  // +1 second margin
+    // reflection padding needs one sample plus half an FFT window
+    size_t min_samples = (size_t) hparams.audio_n_fft / 2 + 1;
     if (n_samples < min_samples) {
         smpl.resize(min_samples, 0.0f);
         std::memcpy(smpl.data(), samples, n_samples * sizeof(float));

tools/server/server-tools.cpp

@@ -17,12 +17,60 @@
 #include <functional>
 #include <memory>
 
+#if defined(_WIN32)
+#   ifndef NOMINMAX
+#       define NOMINMAX
+#   endif
+#   include <windows.h>
+#endif
+
 namespace fs = std::filesystem;
 
 //
 // internal helpers
 //
 
+#if defined(_WIN32)
+// A chunk can end in the middle of a multi-byte sequence, so the incomplete
+// tail is dropped before validating what precedes it.
+static bool is_utf8_text(const std::string & text) {
+    return is_valid_utf8(text.substr(0, validate_utf8(text)));
+}
+
+// A child process writes its output in the OEM code page, which is not UTF-8
+// on a western Windows install, so accented text reaches the JSON layer as
+// invalid bytes and is replaced there. Text that already decodes as UTF-8 is
+// returned untouched, so a child that emits UTF-8 is never decoded twice.
+// run() spawns without a console, so the console code page does not apply.
+static std::string console_output_to_utf8(const std::string & text) {
+    if (text.empty() || is_utf8_text(text)) {
+        return text;
+    }
+
+    const UINT cp = GetOEMCP();
+
+    // fail rather than emit replacement characters when the code page is wrong
+    const int wide_len = MultiByteToWideChar(cp, MB_ERR_INVALID_CHARS, text.data(), (int) text.size(), nullptr, 0);
+    if (wide_len <= 0) {
+        return text;
+    }
+    std::wstring wide(wide_len, L'\0');
+    MultiByteToWideChar(cp, MB_ERR_INVALID_CHARS, text.data(), (int) text.size(), wide.data(), wide_len);
+
+    const int utf8_len = WideCharToMultiByte(CP_UTF8, 0, wide.data(), wide_len, nullptr, 0, nullptr, nullptr);
+    if (utf8_len <= 0) {
+        return text;
+    }
+    std::string utf8(utf8_len, '\0');
+    WideCharToMultiByte(CP_UTF8, 0, wide.data(), wide_len, utf8.data(), utf8_len, nullptr, nullptr);
+    return utf8;
+}
+#else
+static std::string console_output_to_utf8(const std::string & text) {
+    return text;
+}
+#endif
+
 json server_tool::to_json() const {
     return {
         {"display_name", display_name},
@@ -246,14 +294,14 @@ public:
                     size_t len = strlen(buf);
                     if (output.size() + len <= max_output) {
                         output.append(buf, len);
-                        if (on_chunk && !on_chunk(std::string(buf, len))) {
+                        if (on_chunk && !on_chunk(console_output_to_utf8(std::string(buf, len)))) {
                             proc.terminate();
                             break;
                         }
                     } else {
                         size_t remaining = max_output - output.size();
                         output.append(buf, remaining);
-                        if (on_chunk && remaining > 0) on_chunk(std::string(buf, remaining));
+                        if (on_chunk && remaining > 0) on_chunk(console_output_to_utf8(std::string(buf, remaining)));
                         truncated = true;
                     }
                 }
@@ -267,7 +315,7 @@ public:
 
         res.exit_code = proc.join();
 
-        res.output    = output;
+        res.output    = console_output_to_utf8(output);
         res.timed_out = timed_out.load();
         if (truncated) {
             res.output += "\n[output truncated]";

tools/ui/.storybook/main.ts

@@ -11,7 +11,8 @@ const config: StorybookConfig = {
 		'@chromatic-com/storybook',
 		'@storybook/addon-vitest',
 		'@storybook/addon-a11y',
-		'@storybook/addon-docs'
+		'@storybook/addon-docs',
+		'@storybook/addon-mcp'
 	],
 	framework: '@storybook/sveltekit',
 	viteFinal: async (config) => {

tools/ui/.storybook/vitest.setup.ts

@@ -1,12 +0,0 @@
-import * as a11yAddonAnnotations from '@storybook/addon-a11y/preview';
-import { setProjectAnnotations } from '@storybook/sveltekit';
-import * as previewAnnotations from './preview';
-import { beforeAll } from 'vitest';
-
-const project = setProjectAnnotations([a11yAddonAnnotations, previewAnnotations]);
-
-beforeAll(async () => {
-	if (project.beforeAll) {
-		await project.beforeAll();
-	}
-});

tools/ui/README.md

@@ -89,7 +89,7 @@ Llama UI supports two server operation modes:
 
 ```bash
 cd tools/ui
-npm install
+npm ci
 ```
 
 ### 2. Start llama-server

tools/ui/package.json

@@ -27,20 +27,20 @@
 		"cleanup": "rm -rf .svelte-kit build node_modules test-results dist dev-dist debug-storybook.log static/pwa-*.png static/maskable-icon-*.png static/apple-touch-icon-*.png static/apple-splash-*.png static/favicon*.ico"
 	},
 	"devDependencies": {
-		"@chromatic-com/storybook": "5.0.0",
+		"@chromatic-com/storybook": "5.2.1",
 		"@eslint/compat": "1.4.1",
 		"@eslint/js": "9.39.2",
 		"@internationalized/date": "3.12.2",
 		"@lucide/svelte": "1.25.0",
-		"@modelcontextprotocol/sdk": "1.26.0",
+		"@modelcontextprotocol/sdk": "1.30.0",
 		"@playwright/test": "1.56.1",
-		"@storybook/addon-a11y": "10.2.4",
-		"@storybook/addon-docs": "10.2.4",
-		"@storybook/addon-svelte-csf": "5.0.10",
-		"@storybook/addon-vitest": "10.2.4",
-		"@storybook/sveltekit": "10.2.4",
+		"@storybook/addon-a11y": "10.5.6",
+		"@storybook/addon-docs": "10.5.6",
+		"@storybook/addon-svelte-csf": "5.1.2",
+		"@storybook/addon-vitest": "10.5.6",
+		"@storybook/sveltekit": "10.5.6",
 		"@sveltejs/adapter-static": "3.0.10",
-		"@sveltejs/kit": "2.60.1",
+		"@sveltejs/kit": "2.70.2",
 		"@sveltejs/vite-plugin-svelte": "6.2.1",
 		"@tailwindcss/forms": "0.5.10",
 		"@tailwindcss/typography": "0.5.16",
@@ -48,16 +48,16 @@
 		"@types/node": "24.13.0",
 		"@vite-pwa/assets-generator": "1.0.2",
 		"@vite-pwa/sveltekit": "1.1.0",
-		"@vitest/browser": "4.1.8",
-		"@vitest/browser-playwright": "4.1.8",
-		"@vitest/coverage-v8": "4.1.8",
+		"@vitest/browser": "4.1.10",
+		"@vitest/browser-playwright": "4.1.10",
+		"@vitest/coverage-v8": "4.1.10",
 		"bits-ui": "2.18.1",
 		"clsx": "2.1.1",
 		"dexie": "4.4.3",
-		"dompurify": "3.4.11",
+		"dompurify": "3.4.13",
 		"eslint": "9.39.4",
 		"eslint-config-prettier": "10.1.8",
-		"eslint-plugin-storybook": "10.4.2",
+		"eslint-plugin-storybook": "10.5.6",
 		"eslint-plugin-svelte": "3.19.0",
 		"fflate": "0.8.3",
 		"globals": "16.5.0",
@@ -82,7 +82,7 @@
 		"remark-math": "6.0.0",
 		"remark-rehype": "11.1.2",
 		"sass": "1.100.0",
-		"storybook": "10.4.2",
+		"storybook": "10.5.6",
 		"svelte": "5.56.1",
 		"svelte-check": "4.6.0",
 		"svelte-sonner": "1.1.1",
@@ -95,13 +95,16 @@
 		"unified": "11.0.5",
 		"unist-util-visit": "5.1.0",
 		"uuid": "13.0.2",
-		"vite": "7.3.5",
+		"vite": "7.3.6",
 		"vite-plugin-devtools-json": "0.2.1",
-		"vitest": "4.1.8",
+		"vitest": "4.1.10",
 		"vitest-browser-svelte": "2.1.1",
-		"workbox-window": "7.4.1"
+		"workbox-window": "7.4.1",
+		"@storybook/addon-mcp": "0.7.0"
 	},
 	"overrides": {
-		"cookie": "1.1.1"
+		"cookie": "1.1.1",
+		"sharp": "0.35.3",
+		"valibot": "1.4.2"
 	}
 }

tools/ui/scripts/dev.sh

@@ -14,7 +14,7 @@ cd ../../
 # Ensure node_modules are installed
 if [ ! -d "tools/ui/node_modules" ]; then
     echo "📦 Installing npm dependencies..."
-    cd tools/ui && npm install && cd ../../
+    cd tools/ui && npm ci && cd ../../
 fi
 
 # Check and install git hooks if missing

tools/ui/scripts/git-hooks/pre-commit.sh

@@ -14,7 +14,7 @@ cd "$REPO_ROOT/tools/ui"
 
 # Check that node_modules exists
 if [ ! -d "node_modules" ]; then
-    echo "❌ node_modules not found. Run 'npm install' first."
+    echo "❌ node_modules not found. Run 'npm ci' first."
     exit 1
 fi
 

tools/ui/scripts/git-hooks/pre-push.sh

@@ -30,7 +30,7 @@ cd "$REPO_ROOT/tools/ui"
 
 # Check that node_modules exists
 if [ ! -d "node_modules" ]; then
-    echo "❌ node_modules not found. Run 'npm install' first."
+    echo "❌ node_modules not found. Run 'npm ci' first."
     exit 1
 fi
 

tools/ui/src/lib/constants/settings-registry.ts

@@ -198,7 +198,7 @@ const SETTINGS_REGISTRY: Record<string, SettingsSectionEntry> = {
 				key: SETTINGS_KEYS.SHOW_MESSAGE_STATS,
 				label: 'Show message generation statistics',
 				help: 'Display generation statistics (tokens/second, token count, duration) below each assistant message.',
-				defaultValue: false,
+				defaultValue: true,
 				type: SettingsFieldType.CHECKBOX,
 				section: SETTINGS_SECTION_SLUGS.DISPLAY
 			},

tools/ui/vite.config.ts

@@ -76,8 +76,7 @@ export default defineConfig({
 				extends: './vite.config.ts',
 				test: {
 					name: 'ui',
-					browser: { ...browserBaseConfig, instances: [{ browser: 'chromium', headless: true }] },
-					setupFiles: ['./.storybook/vitest.setup.ts']
+					browser: { ...browserBaseConfig, instances: [{ browser: 'chromium', headless: true }] }
 				},
 				plugins: [
 					storybookTest({