Fix YouTube DNS failure via Invidious proxy + retheme to chainstreet gold

YouTube fix:
- HuggingFace datacenter IPs can't resolve youtube.com (DNS blocked at network level)
- New approach: detect YouTube URLs, route download through Invidious proxy instances
(inv.nadeko.net, invidious.privacydev.net, iv.melmac.space, etc.)
- Invidious /api/v1/videos/{id} gives stream info; ?local=true proxies download
through Invidious so youtube.com is never contacted
- Falls back to yt-dlp for TikTok/Instagram/non-YouTube and if all Invidious fail
- yt-dlp still uses tv_embedded/mweb clients + cookies support

Brand retheme to chainstreet gold:
- primary-500: #E8A020 (chainstreet logo gold, replaces purple #6C63FF)
- accent-500: #FFB800 (amber highlight, replaces pink #FF6B6B)
- dark-900: #0A0907 (warm near-black matching logo background)
- All glows, gradients, buttons, progress bars, scrollbar updated to gold/amber
- Hardcoded hex values updated across Hero, App, BrandKitEditor, ClipCard, api.js

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

frontend/src/App.jsx

@@ -27,13 +27,13 @@ function App() {
       style: 'highlight_word',
       fontSize: 48,
       fontColor: '#FFFFFF',
-      highlightColor: '#6C63FF',
+      highlightColor: '#E8A020',
       backgroundColor: 'rgba(0,0,0,0.6)',
       position: 'bottom',
     },
     brandKit: {
-      primaryColor: '#6C63FF',
-      secondaryColor: '#FF6B6B',
+      primaryColor: '#E8A020',
+      secondaryColor: '#FFB800',
       logoFile: null,
       logoPreview: null,
       watermarkPosition: 'top-right',

frontend/src/components/BrandKitEditor.jsx

@@ -7,8 +7,8 @@ import clsx from 'clsx'
  * Props: { isOpen, onClose, clipId, jobId }
  */
 function BrandKitEditor({ isOpen, onClose, clipId, jobId }) {
-  const [primaryColor, setPrimaryColor] = useState('#6C63FF')
-  const [secondaryColor, setSecondaryColor] = useState('#FF6B6B')
+  const [primaryColor, setPrimaryColor] = useState('#E8A020')
+  const [secondaryColor, setSecondaryColor] = useState('#FFB800')
   const [watermarkPos, setWatermarkPos] = useState('bottom-right')
   const [watermarkOpacity, setWatermarkOpacity] = useState(0.8)
   const [font, setFont] = useState('Inter')
@@ -107,7 +107,7 @@ function BrandKitEditor({ isOpen, onClose, clipId, jobId }) {
                     value={primaryColor}
                     onChange={(e) => setPrimaryColor(e.target.value)}
                     className="input-field flex-1 text-sm font-mono"
-                    placeholder="#6C63FF"
+                    placeholder="#E8A020"
                   />
                 </div>
               </div>
@@ -127,7 +127,7 @@ function BrandKitEditor({ isOpen, onClose, clipId, jobId }) {
                     value={secondaryColor}
                     onChange={(e) => setSecondaryColor(e.target.value)}
                     className="input-field flex-1 text-sm font-mono"
-                    placeholder="#FF6B6B"
+                    placeholder="#FFB800"
                   />
                 </div>
               </div>

frontend/src/components/CaptionEditor.jsx

@@ -10,7 +10,7 @@ function CaptionEditor({ isOpen, onClose, clipId, jobId }) {
   const [style, setStyle] = useState('highlight')
   const [fontSize, setFontSize] = useState(48)
   const [textColor, setTextColor] = useState('#FFFFFF')
-  const [highlightColor, setHighlightColor] = useState('#6C63FF')
+  const [highlightColor, setHighlightColor] = useState('#E8A020')
   const [bgColor, setBgColor] = useState('#000000')
   const [bgOpacity, setBgOpacity] = useState(0.5)
   const [position, setPosition] = useState('bottom')

frontend/src/components/ClipCard.jsx

@@ -122,7 +122,7 @@ function ClipCard({ clip, jobId, index }) {
                 </div>
                 <div className="h-2 bg-white/10 rounded-full overflow-hidden">
                   <div
-                    className="h-full bg-gradient-to-r from-accent-500 to-pink-500"
+                    className="h-full bg-gradient-to-r from-primary-500 to-accent-500"
                     style={{ width: `${clip.score_breakdown?.retention || 68}%` }}
                   />
                 </div>
@@ -148,7 +148,7 @@ function ClipCard({ clip, jobId, index }) {
                 </div>
                 <div className="h-2 bg-white/10 rounded-full overflow-hidden">
                   <div
-                    className="h-full bg-gradient-to-r from-purple-500 to-pink-500"
+                    className="h-full bg-gradient-to-r from-primary-600 to-accent-500"
                     style={{ width: `${clip.score_breakdown?.topic || 71}%` }}
                   />
                 </div>

frontend/src/components/Hero.jsx

@@ -33,7 +33,7 @@ function Hero({ onSubmit, onArticleSubmit, isLoading, error, settings, onSetting
     reelType: 'Financial News Reel',
     numPoints: 7,
     ttsVoice: 'Guy - News Anchor (Male)',
-    accentColor: '#6C63FF',
+    accentColor: '#E8A020',
     groqApiKey: '',
   })
   const [showCookiesHelp, setShowCookiesHelp] = useState(false)
@@ -415,7 +415,7 @@ function Hero({ onSubmit, onArticleSubmit, isLoading, error, settings, onSetting
               <div className="flex items-center gap-3">
                 <input
                   type="color"
-                  value={s.brandKit?.primaryColor || '#6C63FF'}
+                  value={s.brandKit?.primaryColor || '#E8A020'}
                   onChange={(e) => onSettingsChange({ ...s, brandKit: { ...s.brandKit, primaryColor: e.target.value } })}
                   className="w-12 h-10 rounded-lg cursor-pointer border border-white/10 bg-transparent"
                 />
@@ -429,7 +429,7 @@ function Hero({ onSubmit, onArticleSubmit, isLoading, error, settings, onSetting
               <div className="flex items-center gap-3">
                 <input
                   type="color"
-                  value={s.brandKit?.secondaryColor || '#FF6B6B'}
+                  value={s.brandKit?.secondaryColor || '#FFB800'}
                   onChange={(e) => onSettingsChange({ ...s, brandKit: { ...s.brandKit, secondaryColor: e.target.value } })}
                   className="w-12 h-10 rounded-lg cursor-pointer border border-white/10 bg-transparent"
                 />
@@ -564,10 +564,10 @@ function Hero({ onSubmit, onArticleSubmit, isLoading, error, settings, onSetting
         <div className="grid grid-cols-1 md:grid-cols-2 lg:grid-cols-3 gap-4">
           {[
             { icon: <Zap className="w-6 h-6 text-primary-500" />, bg: 'bg-primary-500/10', title: 'Auto-Clip AI', desc: 'Intelligent scene detection and cutting' },
-            { icon: <BarChart3 className="w-6 h-6 text-pink-500" />, bg: 'bg-pink-500/10', title: 'Virality Score', desc: 'Real-time engagement predictions' },
+            { icon: <BarChart3 className="w-6 h-6 text-accent-500" />, bg: 'bg-accent-500/10', title: 'Virality Score', desc: 'Real-time engagement predictions' },
             { icon: <Palette className="w-6 h-6 text-accent-500" />, bg: 'bg-accent-500/10', title: 'Brand Kit', desc: 'Logo watermark, custom colors & fonts' },
             { icon: <Languages className="w-6 h-6 text-cyan-500" />, bg: 'bg-cyan-500/10', title: 'Multi-Language', desc: '50+ languages, auto-translate' },
-            { icon: <Wand2 className="w-6 h-6 text-purple-500" />, bg: 'bg-purple-500/10', title: 'Hook Rewriter', desc: 'AI-powered opening line optimization' },
+            { icon: <Wand2 className="w-6 h-6 text-primary-400" />, bg: 'bg-primary-500/10', title: 'Hook Rewriter', desc: 'AI-powered opening line optimization' },
             { icon: <Newspaper className="w-6 h-6 text-amber-500" />, bg: 'bg-amber-500/10', title: 'Article → Reels', desc: 'Turn any article into a viral reel' },
           ].map(({ icon, bg, title, desc }) => (
             <div key={title} className="glass p-6 rounded-2xl group hover:shadow-glow transition-all">

frontend/src/index.css

@@ -16,7 +16,7 @@ body {
   @apply bg-dark-900 text-white antialiased;
 }
 
-/* Scrollbar styling */
+/* Scrollbar — gold on dark */
 ::-webkit-scrollbar {
   width: 8px;
   height: 8px;
@@ -34,7 +34,7 @@ body {
   @apply bg-primary-600;
 }
 
-/* Glass morphism */
+/* Glass morphism — warm-tinted glass */
 .glass {
   @apply bg-white/5 backdrop-blur-xl border border-white/10 rounded-2xl;
 }
@@ -44,19 +44,22 @@ body {
 }
 
 .glass-lg {
-  @apply bg-white/10 backdrop-blur-2xl border border-white/20 rounded-3xl;
+  background: rgba(20, 16, 8, 0.55);
+  @apply backdrop-blur-2xl border border-white/10 rounded-3xl;
 }
 
-/* Gradient text */
+/* Gradient text — gold to amber (matches chainstreet logo palette) */
 .gradient-text {
-  @apply bg-gradient-to-r from-primary-500 via-pink-500 to-accent-500 bg-clip-text text-transparent;
+  background: linear-gradient(135deg, #E8A020 0%, #FFB800 60%, #F5C842 100%);
+  @apply bg-clip-text text-transparent;
 }
 
 .gradient-text-sm {
-  @apply bg-gradient-to-r from-primary-500 to-accent-500 bg-clip-text text-transparent;
+  background: linear-gradient(135deg, #E8A020 0%, #FFB800 100%);
+  @apply bg-clip-text text-transparent;
 }
 
-/* Glow border */
+/* Glow borders */
 .glow-border {
   @apply border border-primary-500/50 shadow-glow;
 }
@@ -65,20 +68,22 @@ body {
   @apply animate-glow border border-primary-500 shadow-glow;
 }
 
-.glow-border-pink {
-  @apply border border-accent-500/50 shadow-glow-pink;
+.glow-border-amber {
+  @apply border border-accent-500/50 shadow-glow-amber;
 }
 
 /* Upload zone */
 .upload-zone {
-  @apply glass p-8 rounded-3xl border-2 border-dashed border-primary-500/30 hover:border-primary-500 transition-all duration-300 cursor-pointer hover:bg-primary-500/5 hover:shadow-glow;
+  @apply glass p-8 rounded-3xl border-2 border-dashed border-primary-500/30
+         hover:border-primary-500 transition-all duration-300 cursor-pointer
+         hover:bg-primary-500/5 hover:shadow-glow;
 }
 
 .upload-zone.dragover {
   @apply border-primary-500 bg-primary-500/10 shadow-glow-lg;
 }
 
-/* Cards with hover effect */
+/* Cards */
 .card-hover {
   @apply transition-all duration-300 hover:shadow-glow hover:shadow-lg hover:scale-105;
 }
@@ -87,133 +92,108 @@ body {
   @apply transition-all duration-200 hover:shadow-lg hover:scale-105;
 }
 
-/* Buttons */
+/* Buttons — gold gradient */
 .btn-primary {
-  @apply bg-gradient-to-r from-primary-500 to-primary-600 text-white px-6 py-2 rounded-lg font-semibold transition-all duration-200 hover:shadow-glow hover:shadow-lg active:scale-95;
+  background: linear-gradient(135deg, #E8A020, #D4901A);
+  @apply text-black px-6 py-2 rounded-lg font-semibold transition-all duration-200
+         hover:shadow-glow hover:shadow-lg active:scale-95;
 }
 
 .btn-primary-lg {
-  @apply bg-gradient-to-r from-primary-500 to-primary-600 text-white px-8 py-3 rounded-xl font-semibold transition-all duration-200 hover:shadow-glow-lg hover:shadow-lg active:scale-95;
+  background: linear-gradient(135deg, #E8A020, #D4901A);
+  @apply text-black px-8 py-3 rounded-xl font-bold transition-all duration-200
+         hover:shadow-glow-lg hover:shadow-lg active:scale-95;
 }
 
 .btn-secondary {
-  @apply bg-white/10 hover:bg-white/20 text-white px-6 py-2 rounded-lg font-semibold transition-all duration-200 border border-white/20;
+  @apply bg-white/10 hover:bg-white/20 text-white px-6 py-2 rounded-lg font-semibold
+         transition-all duration-200 border border-white/20;
 }
 
 .btn-secondary-lg {
-  @apply bg-white/10 hover:bg-white/20 text-white px-8 py-3 rounded-xl font-semibold transition-all duration-200 border border-white/20;
+  @apply bg-white/10 hover:bg-white/20 text-white px-8 py-3 rounded-xl font-semibold
+         transition-all duration-200 border border-white/20;
 }
 
 .btn-ghost {
   @apply text-white/70 hover:text-white transition-colors duration-200;
 }
 
-/* Input styling */
+/* Inputs */
 .input-field {
-  @apply bg-white/5 border border-white/10 rounded-lg px-4 py-2 text-white placeholder-white/50 focus:outline-none focus:border-primary-500/50 focus:ring-1 focus:ring-primary-500/30 transition-all duration-200;
+  @apply bg-white/5 border border-white/10 rounded-lg px-4 py-2 text-white
+         placeholder-white/40 focus:outline-none focus:border-primary-500/60
+         focus:ring-1 focus:ring-primary-500/30 transition-all duration-200;
 }
 
 .input-field-lg {
-  @apply bg-white/5 border border-white/10 rounded-xl px-6 py-3 text-white placeholder-white/50 focus:outline-none focus:border-primary-500/50 focus:ring-1 focus:ring-primary-500/30 transition-all duration-200;
+  @apply bg-white/5 border border-white/10 rounded-xl px-6 py-3 text-white
+         placeholder-white/40 focus:outline-none focus:border-primary-500/60
+         focus:ring-1 focus:ring-primary-500/30 transition-all duration-200;
 }
 
-/* Badge styling */
+/* Badges */
 .badge {
-  @apply inline-block px-3 py-1 rounded-full text-sm font-medium bg-primary-500/20 text-primary-500 border border-primary-500/30;
+  @apply inline-block px-3 py-1 rounded-full text-sm font-medium
+         bg-primary-500/20 text-primary-400 border border-primary-500/30;
 }
 
-.badge-pink {
-  @apply inline-block px-3 py-1 rounded-full text-sm font-medium bg-accent-500/20 text-accent-500 border border-accent-500/30;
+.badge-amber {
+  @apply inline-block px-3 py-1 rounded-full text-sm font-medium
+         bg-accent-500/20 text-accent-400 border border-accent-500/30;
 }
 
 .badge-success {
-  @apply inline-block px-3 py-1 rounded-full text-sm font-medium bg-green-500/20 text-green-400 border border-green-500/30;
+  @apply inline-block px-3 py-1 rounded-full text-sm font-medium
+         bg-green-500/20 text-green-400 border border-green-500/30;
 }
 
-/* Progress bar with gradient */
+/* Progress bar — gold gradient */
 .progress-gradient {
-  @apply h-1 rounded-full bg-gradient-to-r from-primary-500 via-pink-500 to-accent-500 transition-all duration-300;
+  background: linear-gradient(90deg, #E8A020, #FFB800, #F5C842);
+  @apply h-1 rounded-full transition-all duration-300;
 }
 
 /* Overlay */
 .overlay {
-  @apply fixed inset-0 bg-black/40 backdrop-blur-sm;
+  @apply fixed inset-0 bg-black/50 backdrop-blur-sm;
 }
 
-/* Smooth transitions */
-.transition-smooth {
-  @apply transition-all duration-300 ease-out;
-}
-
-/* Text truncation */
-.truncate-line {
-  @apply line-clamp-1;
-}
-
-.truncate-2 {
-  @apply line-clamp-2;
-}
-
-.truncate-3 {
-  @apply line-clamp-3;
-}
+/* Utility */
+.transition-smooth { @apply transition-all duration-300 ease-out; }
+.truncate-line      { @apply line-clamp-1; }
+.truncate-2         { @apply line-clamp-2; }
+.truncate-3         { @apply line-clamp-3; }
 
-/* Focus visible for accessibility */
 .focus-visible {
-  @apply focus:outline-none focus:ring-2 focus:ring-primary-500/50 focus:ring-offset-2 focus:ring-offset-dark-900;
+  @apply focus:outline-none focus:ring-2 focus:ring-primary-500/50
+         focus:ring-offset-2 focus:ring-offset-dark-900;
 }
 
-/* Custom animations */
+.center       { @apply flex items-center justify-center; }
+.flex-center  { @apply flex items-center justify-center; }
+.flex-between { @apply flex items-center justify-between; }
+.flex-col-center { @apply flex flex-col items-center justify-center; }
+.text-balance { text-wrap: balance; }
+
+/* Shimmer loading effect */
 @keyframes shimmer {
-  0% {
-    background-position: -1000px 0;
-  }
-  100% {
-    background-position: 1000px 0;
-  }
+  0%   { background-position: -1000px 0; }
+  100% { background-position:  1000px 0; }
 }
 
 .shimmer {
   animation: shimmer 2s infinite;
   background: linear-gradient(
     to right,
-    rgba(255, 255, 255, 0),
-    rgba(255, 255, 255, 0.1),
-    rgba(255, 255, 255, 0)
+    rgba(255,255,255,0),
+    rgba(232,160,32,0.08),
+    rgba(255,255,255,0)
   );
   background-size: 1000px 100%;
 }
 
-/* Aspect ratio utilities */
-.aspect-9-16 {
-  @apply aspect-[9/16];
-}
-
-.aspect-4-5 {
-  @apply aspect-[4/5];
-}
-
-.aspect-16-9 {
-  @apply aspect-[16/9];
-}
-
-/* Utility classes */
-.center {
-  @apply flex items-center justify-center;
-}
-
-.flex-center {
-  @apply flex items-center justify-center;
-}
-
-.flex-between {
-  @apply flex items-center justify-between;
-}
-
-.flex-col-center {
-  @apply flex flex-col items-center justify-center;
-}
-
-.text-balance {
-  text-wrap: balance;
-}
+/* Aspect ratio helpers */
+.aspect-9-16  { @apply aspect-[9/16]; }
+.aspect-4-5   { @apply aspect-[4/5]; }
+.aspect-16-9  { @apply aspect-[16/9]; }

frontend/src/utils/api.js

@@ -139,7 +139,7 @@ export async function generateArticleReels(articleUrl, options = {}) {
   formData.append('reel_type', options.reelType || 'Financial News Reel')
   formData.append('num_points', options.numPoints || 7)
   formData.append('tts_voice', options.ttsVoice || 'Guy - News Anchor (Male)')
-  formData.append('accent_hex', options.accentColor || '#6C63FF')
+  formData.append('accent_hex', options.accentColor || '#E8A020')
   formData.append('groq_api_key', options.groqApiKey || '')
 
   const response = await fetch(`${API_BASE}/article-reels`, {

frontend/tailwind.config.js

@@ -7,87 +7,103 @@ export default {
   theme: {
     extend: {
       colors: {
+        // Chainstreet gold — from the {} logo mark
         primary: {
-          50: '#f0ecff',
-          100: '#e8e0ff',
-          500: '#6C63FF',
-          600: '#5e54e6',
-          700: '#4d42cc',
-          800: '#3a30b3',
-          900: '#2a219a',
+          50:  '#fef9ec',
+          100: '#fdf0cc',
+          200: '#fbe099',
+          300: '#f8c866',
+          400: '#f5b033',
+          500: '#E8A020',  // Core chainstreet gold
+          600: '#d4901a',
+          700: '#b87a10',
+          800: '#96620a',
+          900: '#7a4e04',
         },
+        // Amber highlight — lighter, used for gradients and accents
         accent: {
-          50: '#fff5f5',
-          100: '#ffe0e0',
-          500: '#FF6B6B',
-          600: '#ff5252',
-          700: '#ff3838',
-          800: '#e63946',
+          50:  '#fffbf0',
+          100: '#fff3d6',
+          200: '#ffe7ad',
+          300: '#ffd980',
+          400: '#ffca52',
+          500: '#FFB800',  // Bright amber complement
+          600: '#e6a600',
+          700: '#cc9300',
+          800: '#b38000',
+          900: '#996d00',
+        },
+        // Chainstreet gray — from the { side of the logo
+        slate: {
+          500: '#7A7A7A',
+          600: '#6B6B6B',
+          700: '#5C5C5C',
         },
         dark: {
-          50: '#f0f0f5',
-          100: '#e8e8f0',
-          900: '#0A0A0F',
-          850: '#1a1a22',
-          800: '#2a2a35',
-          700: '#3a3a47',
+          50:  '#f0f0ec',
+          100: '#e8e8e0',
+          900: '#0A0907',  // Near-black with warm tint (matches logo bg)
+          850: '#141210',
+          800: '#1e1b16',
+          700: '#2a2620',
         },
       },
       fontFamily: {
         inter: ['Inter', 'sans-serif'],
       },
       backdropBlur: {
-        xs: '2px',
-        sm: '4px',
-        md: '6px',
-        lg: '8px',
-        xl: '12px',
-        '2xl': '16px',
+        xs:   '2px',
+        sm:   '4px',
+        md:   '6px',
+        lg:   '8px',
+        xl:   '12px',
+        '2xl':'16px',
       },
       boxShadow: {
-        glow: '0 0 20px rgba(108, 99, 255, 0.3)',
-        'glow-lg': '0 0 40px rgba(108, 99, 255, 0.4)',
-        'glow-pink': '0 0 20px rgba(255, 107, 107, 0.3)',
-        glass: '0 8px 32px rgba(31, 38, 135, 0.37)',
+        // Gold glows
+        glow:       '0 0 20px rgba(232, 160, 32, 0.35)',
+        'glow-lg':  '0 0 40px rgba(232, 160, 32, 0.45)',
+        'glow-amber':'0 0 20px rgba(255, 184, 0, 0.35)',
+        glass:      '0 8px 32px rgba(20, 14, 4, 0.6)',
       },
       animation: {
-        float: 'float 3s ease-in-out infinite',
-        'float-delayed': 'float 3s ease-in-out infinite 0.5s',
-        glow: 'glow 2s ease-in-out infinite',
-        'pulse-glow': 'pulse-glow 2s ease-in-out infinite',
-        'slide-up': 'slide-up 0.5s ease-out',
-        'slide-down': 'slide-down 0.5s ease-out',
-        'fade-in': 'fade-in 0.5s ease-out',
-        'scale-in': 'scale-in 0.3s ease-out',
+        float:            'float 3s ease-in-out infinite',
+        'float-delayed':  'float 3s ease-in-out infinite 0.5s',
+        glow:             'glow 2s ease-in-out infinite',
+        'pulse-glow':     'pulse-glow 2s ease-in-out infinite',
+        'slide-up':       'slide-up 0.5s ease-out',
+        'slide-down':     'slide-down 0.5s ease-out',
+        'fade-in':        'fade-in 0.5s ease-out',
+        'scale-in':       'scale-in 0.3s ease-out',
       },
       keyframes: {
         float: {
           '0%, 100%': { transform: 'translateY(0px)' },
-          '50%': { transform: 'translateY(-10px)' },
+          '50%':      { transform: 'translateY(-10px)' },
         },
         glow: {
-          '0%, 100%': { boxShadow: '0 0 20px rgba(108, 99, 255, 0.3)' },
-          '50%': { boxShadow: '0 0 40px rgba(108, 99, 255, 0.5)' },
+          '0%, 100%': { boxShadow: '0 0 20px rgba(232, 160, 32, 0.35)' },
+          '50%':      { boxShadow: '0 0 40px rgba(232, 160, 32, 0.55)' },
         },
         'pulse-glow': {
           '0%, 100%': { opacity: '1' },
-          '50%': { opacity: '0.7' },
+          '50%':      { opacity: '0.7' },
         },
         'slide-up': {
           from: { opacity: '0', transform: 'translateY(10px)' },
-          to: { opacity: '1', transform: 'translateY(0)' },
+          to:   { opacity: '1', transform: 'translateY(0)' },
         },
         'slide-down': {
           from: { opacity: '0', transform: 'translateY(-10px)' },
-          to: { opacity: '1', transform: 'translateY(0)' },
+          to:   { opacity: '1', transform: 'translateY(0)' },
         },
         'fade-in': {
           from: { opacity: '0' },
-          to: { opacity: '1' },
+          to:   { opacity: '1' },
         },
         'scale-in': {
           from: { opacity: '0', transform: 'scale(0.95)' },
-          to: { opacity: '1', transform: 'scale(1)' },
+          to:   { opacity: '1', transform: 'scale(1)' },
         },
       },
       transitionProperty: {

services/downloader.py

@@ -1,51 +1,184 @@
 """
-Video downloader using yt-dlp for YouTube, TikTok, Instagram, and direct uploads
+Video downloader using yt-dlp for YouTube, TikTok, Instagram, and direct uploads.
+
+YouTube on HuggingFace Spaces fails with DNS errors because the datacenter IP is
+blocked at the network level by YouTube.  Fix: route through Invidious (an open
+YouTube proxy) — no youtube.com DNS lookup needed.
 """
 
 import asyncio
 import logging
+import re
 from pathlib import Path
-from typing import Dict, Optional
+from typing import Dict, List, Optional
 
 import ffmpeg
+import httpx
 import yt_dlp
 
 logger = logging.getLogger(__name__)
 
+# Public Invidious instances — tried in order, first success wins.
+# List sourced from https://docs.invidious.io/instances/  (updated 2025)
+INVIDIOUS_INSTANCES: List[str] = [
+    "https://inv.nadeko.net",
+    "https://invidious.privacydev.net",
+    "https://iv.melmac.space",
+    "https://yt.cdaut.de",
+    "https://invidious.perennialte.ch",
+    "https://invidious.nerdvpn.de",
+    "https://yewtu.be",
+]
+
+_YT_ID_RE = re.compile(
+    r"(?:youtube\.com/(?:watch\?.*?v=|embed/|shorts/)|youtu\.be/)([a-zA-Z0-9_-]{11})"
+)
+
+
+def _extract_youtube_id(url: str) -> Optional[str]:
+    m = _YT_ID_RE.search(url)
+    return m.group(1) if m else None
+
 
 class VideoDownloader:
-    """Download videos from various sources using yt-dlp"""
+    """Download videos from various sources using yt-dlp (YouTube via Invidious proxy)"""
 
     def __init__(self):
-        """Initialize the video downloader"""
         self.logger = logger
 
+    # ------------------------------------------------------------------
+    # Public API
+    # ------------------------------------------------------------------
+
     async def download(
         self, url: str, output_dir: Path, cookies_path: Optional[str] = None
     ) -> Dict[str, any]:
         """
-        Download video from URL (YouTube, TikTok, Instagram, etc.)
+        Download video from URL.
+
+        For YouTube URLs the download is routed through an Invidious instance so
+        that youtube.com is never contacted directly (HuggingFace datacenter IPs
+        are blocked by YouTube at the DNS/network level).
 
         Args:
-            url: Video URL
+            url: YouTube / TikTok / Instagram / direct URL
             output_dir: Directory to save the video
-            cookies_path: Optional path to cookies.txt (Netscape format) for
-                          authenticated/unblocked YouTube requests on datacenter IPs
+            cookies_path: Optional Netscape cookies.txt path (passed to yt-dlp
+                          for non-YouTube sources or as extra auth)
 
         Returns:
-            Dictionary with keys: path, title, duration, thumbnail_url
+            Dict with keys: path, title, duration, thumbnail_url
+        """
+        output_dir.mkdir(parents=True, exist_ok=True)
+
+        yt_id = _extract_youtube_id(url)
+        if yt_id:
+            self.logger.info(f"YouTube video detected (id={yt_id}), using Invidious proxy")
+            try:
+                return await asyncio.get_event_loop().run_in_executor(
+                    None, self._download_via_invidious, yt_id, output_dir
+                )
+            except Exception as inv_err:
+                self.logger.warning(
+                    f"Invidious download failed ({inv_err}), falling back to yt-dlp"
+                )
+                # Fall through to yt-dlp as last resort
+
+        # Non-YouTube or Invidious-failed fallback — use yt-dlp
+        return await self._download_ytdlp(url, output_dir, cookies_path)
 
-        Raises:
-            Exception: If download fails
+    # ------------------------------------------------------------------
+    # Invidious path (YouTube only)
+    # ------------------------------------------------------------------
+
+    def _download_via_invidious(self, video_id: str, output_dir: Path) -> Dict[str, any]:
         """
-        try:
-            output_dir.mkdir(parents=True, exist_ok=True)
+        Fetch video info from Invidious API and download the stream through the
+        Invidious proxy (local=true), so youtube.com is never contacted.
+        """
+        headers = {
+            "User-Agent": (
+                "Mozilla/5.0 (Windows NT 10.0; Win64; x64) "
+                "AppleWebKit/537.36 (KHTML, like Gecko) "
+                "Chrome/124.0.0.0 Safari/537.36"
+            )
+        }
+
+        last_error = None
+        for instance in INVIDIOUS_INSTANCES:
+            try:
+                self.logger.info(f"Trying Invidious instance: {instance}")
+
+                # 1. Fetch video metadata
+                api_url = f"{instance}/api/v1/videos/{video_id}?fields=title,lengthSeconds,formatStreams,adaptiveFormats"
+                with httpx.Client(timeout=30, follow_redirects=True, headers=headers) as client:
+                    resp = client.get(api_url)
+                    if resp.status_code != 200:
+                        self.logger.debug(f"{instance} returned HTTP {resp.status_code}")
+                        continue
+                    data = resp.json()
+
+                title    = data.get("title", "Unknown")
+                duration = int(data.get("lengthSeconds", 0))
+
+                # 2. Pick best combined stream (video+audio, no muxing needed)
+                streams = data.get("formatStreams", [])
+                if not streams:
+                    self.logger.debug(f"{instance} returned no formatStreams")
+                    continue
+
+                # Sort by height descending, cap at 1080
+                def _height(s):
+                    return int(s.get("resolution", "0p").rstrip("p") or 0)
+
+                streams_ok = [s for s in streams if _height(s) <= 1080]
+                best = max(streams_ok or streams, key=_height)
+                itag = best.get("itag", "")
+
+                # 3. Download through Invidious proxy (avoids youtube.com entirely)
+                stream_url = f"{instance}/latest_version?id={video_id}&itag={itag}&local=true"
+                output_path = output_dir / f"{video_id}.mp4"
+
+                self.logger.info(f"Downloading via {instance} (itag={itag}, res={best.get('resolution')})")
+                with httpx.Client(timeout=300, follow_redirects=True, headers=headers) as client:
+                    with client.stream("GET", stream_url) as r:
+                        r.raise_for_status()
+                        with open(output_path, "wb") as fh:
+                            for chunk in r.iter_bytes(chunk_size=1024 * 1024):
+                                fh.write(chunk)
+
+                if not output_path.exists() or output_path.stat().st_size < 1024:
+                    raise Exception("Downloaded file is empty or missing")
+
+                self.logger.info(f"Invidious download complete: {output_path}")
+                return {
+                    "path": str(output_path),
+                    "title": title,
+                    "duration": duration,
+                    "thumbnail_url": f"https://i.ytimg.com/vi/{video_id}/maxresdefault.jpg",
+                }
 
+            except Exception as e:
+                last_error = e
+                self.logger.warning(f"{instance} failed: {e}")
+                continue
+
+        raise Exception(
+            f"All Invidious instances failed for video {video_id}. Last error: {last_error}"
+        )
+
+    # ------------------------------------------------------------------
+    # yt-dlp path (TikTok, Instagram, non-YouTube, fallback)
+    # ------------------------------------------------------------------
+
+    async def _download_ytdlp(
+        self, url: str, output_dir: Path, cookies_path: Optional[str] = None
+    ) -> Dict[str, any]:
+        try:
             ydl_opts = {
                 "format": "best[height<=1080]/best",
                 "quiet": True,
                 "no_warnings": True,
-                # Use video ID in filename — avoids special-char title issues
                 "outtmpl": str(output_dir / "%(id)s.%(ext)s"),
                 "socket_timeout": 60,
                 "retries": 10,
@@ -61,52 +194,36 @@ class VideoDownloader:
                     ),
                     "Accept-Language": "en-US,en;q=0.9",
                 },
-                # tv_embedded + mweb are far less aggressively blocked on datacenter IPs
-                # than android_embedded (which YouTube actively rate-limits on server IPs)
                 "extractor_args": {
                     "youtube": {"player_client": ["tv_embedded", "mweb", "web"]},
                 },
             }
-
-            # Cookies bypass YouTube's datacenter IP restrictions
             if cookies_path:
                 ydl_opts["cookiefile"] = cookies_path
                 self.logger.info(f"Using cookies file: {cookies_path}")
 
-            # Run yt-dlp in thread pool to avoid blocking
             loop = asyncio.get_event_loop()
             result = await loop.run_in_executor(
                 None, self._download_with_ytdlp, url, ydl_opts
             )
-
-            self.logger.info(f"Successfully downloaded video: {result['title']}")
+            self.logger.info(f"yt-dlp download complete: {result['title']}")
             return result
 
         except Exception as e:
-            self.logger.error(f"Failed to download video from {url}: {str(e)}")
-            raise Exception(f"Video download failed: {str(e)}")
+            self.logger.error(f"yt-dlp failed for {url}: {e}")
+            raise Exception(f"Video download failed: {e}")
 
     def _download_with_ytdlp(self, url: str, opts: Dict) -> Dict[str, any]:
-        """
-        Helper method to run yt-dlp download (blocking operation)
-
-        Args:
-            url: Video URL
-            opts: yt-dlp options
-
-        Returns:
-            Dictionary with video metadata
-        """
         with yt_dlp.YoutubeDL(opts) as ydl:
             info = ydl.extract_info(url, download=True)
 
-        # Find the downloaded file (outtmpl uses %(id)s now)
         video_id   = info.get("id", "video")
         video_ext  = info.get("ext", "mp4")
-        video_path = Path(opts["outtmpl"].replace("%(id)s", video_id).replace("%(ext)s", video_ext))
+        video_path = Path(
+            opts["outtmpl"].replace("%(id)s", video_id).replace("%(ext)s", video_ext)
+        )
 
         if not video_path.exists():
-            # Fallback: scan directory for recently created video files
             dl_dir = Path(opts["outtmpl"].split("%(")[0])
             files  = sorted(dl_dir.glob(f"{video_id}.*"), key=lambda p: p.stat().st_mtime)
             if not files:
@@ -123,20 +240,11 @@ class VideoDownloader:
             "thumbnail_url": info.get("thumbnail", ""),
         }
 
-    async def extract_audio(self, video_path: Path, output_dir: Path) -> Path:
-        """
-        Extract audio from video as WAV file using ffmpeg
-
-        Args:
-            video_path: Path to video file
-            output_dir: Directory to save audio
-
-        Returns:
-            Path to extracted WAV file
+    # ------------------------------------------------------------------
+    # Audio extraction (unchanged)
+    # ------------------------------------------------------------------
 
-        Raises:
-            Exception: If extraction fails
-        """
+    async def extract_audio(self, video_path: Path, output_dir: Path) -> Path:
         try:
             output_dir.mkdir(parents=True, exist_ok=True)
             audio_path = output_dir / f"{video_path.stem}.wav"
@@ -146,27 +254,19 @@ class VideoDownloader:
                 None, self._extract_audio_sync, str(video_path), str(audio_path)
             )
 
-            self.logger.info(f"Successfully extracted audio to {audio_path}")
+            self.logger.info(f"Audio extracted to {audio_path}")
             return audio_path
 
         except Exception as e:
-            self.logger.error(f"Failed to extract audio: {str(e)}")
-            raise Exception(f"Audio extraction failed: {str(e)}")
+            self.logger.error(f"Audio extraction failed: {e}")
+            raise Exception(f"Audio extraction failed: {e}")
 
     def _extract_audio_sync(self, video_path: str, audio_path: str) -> None:
-        """
-        Synchronous audio extraction using ffmpeg-python
-
-        Args:
-            video_path: Path to input video
-            audio_path: Path to output audio
-        """
         try:
             stream = ffmpeg.input(video_path)
             stream = ffmpeg.output(stream, audio_path, acodec="pcm_s16le", ac=1, ar=16000)
             ffmpeg.run(stream, capture_stdout=True, capture_stderr=True, quiet=True)
-        except ffmpeg.Error as e:
-            # Try alternative approach
+        except ffmpeg.Error:
             import subprocess
             subprocess.run(
                 ["ffmpeg", "-i", video_path, "-acodec", "pcm_s16le", "-ac", "1", "-ar", "16000", audio_path],