feat: bump V2 CDN to 0.4.7 with emotion control (FiLM conditioning)

- V2 CDN version 0.4.5 → 0.4.7 (V1 stays at 0.4.5)
- Split VERSION into VERSION_V1/VERSION_V2 in comparison example
- Add V2 emotion control feature (5-dim: neutral, joy, anger, sadness, surprise)
  to homepage, comparison table, README, llms docs, and AI discovery configs
- Update V2 package description and structured data

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

Author: jcggl

.well-known/agent-card.json

@@ -3,7 +3,7 @@
   "name": "AnimaSync",
   "description": "Voice-driven 3D avatar animation engine for the browser. Extracts emotion from speech and generates lip sync, facial expressions, and body motion in real time — entirely client-side via Rust/WASM and ONNX inference.",
   "url": "https://animasync.quasar.ggls.dev/",
-  "version": "0.4.5",
+  "version": "0.4.7",
   "provider": {
     "organization": "GoodGang Labs",
     "url": "https://goodganglabs.com"

.well-known/ai-catalog.json

@@ -53,7 +53,7 @@
       },
       {
         "name": "@goodganglabs/lipsync-wasm-v2",
-        "description": "V2 engine — 52-dim ARKit blendshapes via student distillation. Lightweight alternative.",
+        "description": "V2 engine — 52-dim ARKit blendshapes via student distillation with 5-dim emotion control (FiLM conditioning).",
         "url": "https://www.npmjs.com/package/@goodganglabs/lipsync-wasm-v2"
       }
     ]

README.md

@@ -172,8 +172,9 @@ The production site is available at **[animasync.quasar.ggls.dev](https://animas
 | **Expression generation** | Built-in `IdleExpressionGenerator` (blinks + micro-expressions) | Blink injection in post-process |
 | **Voice activity** | Built-in `VoiceActivityDetector` (body pose switching) | Not included |
 | **ONNX fallback** | Heuristic mode (energy-based) | None (ONNX required) |
+| **Emotion control** | Not included | 5-dim FiLM conditioning (`setEmotion([neutral, joy, anger, sadness, surprise])`) |
 | **Body motion** | VRMA idle/speaking + VAD auto-switch (LoopPingPong, asymmetric crossfade) | VRMA idle/speaking (LoopPingPong, asymmetric crossfade) |
-| **Best for** | Full expression control, custom avatars | Quick integration, lightweight |
+| **Best for** | Full expression control, custom avatars | Emotion-aware lip sync, quick integration |
 
 ---
 

agents.json

@@ -150,7 +150,7 @@
         },
         {
           "action": "init-animasync",
-          "code": "<script type=\"module\">\nconst CDN = 'https://cdn.jsdelivr.net/npm/@goodganglabs/lipsync-wasm-v2@0.4.5';\nconst { LipSyncWasmWrapper } = await import(`${CDN}/lipsync-wasm-wrapper.js`);\nconst lipsync = new LipSyncWasmWrapper({ wasmPath: `${CDN}/lipsync_wasm_v2.js` });\nawait lipsync.init();\n</script>",
+          "code": "<script type=\"module\">\nconst CDN = 'https://cdn.jsdelivr.net/npm/@goodganglabs/lipsync-wasm-v2@0.4.7';\nconst { LipSyncWasmWrapper } = await import(`${CDN}/lipsync-wasm-wrapper.js`);\nconst lipsync = new LipSyncWasmWrapper({ wasmPath: `${CDN}/lipsync_wasm_v2.js` });\nawait lipsync.init();\n</script>",
           "description": "Import and initialize AnimaSync V2 from CDN"
         }
       ]

examples/vanilla-avatar/index.html

@@ -199,7 +199,7 @@ <h2>52 ARKit Blendshapes — V2 Student</h2>
 // No 3D avatar, no Three.js. Pure audio → lip sync data (52-dim).
 // ================================================================
 
-const VERSION = '0.4.5';
+const VERSION = '0.4.7';
 const CDN = `https://cdn.jsdelivr.net/npm/@goodganglabs/lipsync-wasm-v2@${VERSION}`;
 
 // ── All 52 ARKit blendshape channels ──

examples/vanilla-comparison/index.html

@@ -253,9 +253,10 @@ <h1>Anima<span>Sync</span></h1>
 // ================================================================
 // Config
 // ================================================================
-const VERSION = '0.4.5';
-const CDN_V1 = `https://cdn.jsdelivr.net/npm/@goodganglabs/lipsync-wasm-v1@${VERSION}`;
-const CDN_V2 = `https://cdn.jsdelivr.net/npm/@goodganglabs/lipsync-wasm-v2@${VERSION}`;
+const VERSION_V1 = '0.4.5';
+const VERSION_V2 = '0.4.7';
+const CDN_V1 = `https://cdn.jsdelivr.net/npm/@goodganglabs/lipsync-wasm-v1@${VERSION_V1}`;
+const CDN_V2 = `https://cdn.jsdelivr.net/npm/@goodganglabs/lipsync-wasm-v2@${VERSION_V2}`;
 
 const ARKIT_52 = [
   'browDownLeft','browDownRight','browInnerUp','browOuterUpLeft','browOuterUpRight',

index.html

@@ -62,6 +62,7 @@
       "Facial expression generation",
       "Stochastic eye blink animation",
       "VRMA body motion with idle/speaking crossfade",
+      "5-dim emotion control via FiLM conditioning (V2)",
       "VRM avatar support via Three.js",
       "AudioWorklet microphone streaming",
       "Runs entirely in the browser (no server needed)"
@@ -672,10 +673,10 @@ <h2 class="section-title">Install from npm</h2>
         <div class="pkg-name">@goodganglabs/lipsync-wasm-v2</div>
         <span class="pkg-tag pkg-tag-full">Lightweight</span>
       </div>
-      <p class="pkg-desc">Student distillation model &mdash; direct 52-dim ARKit blendshape prediction. Simpler post-processing, faster integration.</p>
+      <p class="pkg-desc">Student distillation model &mdash; direct 52-dim ARKit blendshape prediction with 5-dim emotion control (FiLM conditioning).</p>
       <div class="pkg-meta">
         <div class="pkg-meta-item">Output: <span>52-dim</span> ARKit</div>
-        <div class="pkg-meta-item">Post: <span>crisp_mouth</span> + fade + blink</div>
+        <div class="pkg-meta-item">Emotion: <span>5-dim</span> FiLM conditioning</div>
         <div class="pkg-meta-item">Peer: <span>onnxruntime-web</span></div>
       </div>
       <div class="pkg-install"><span class="dollar">$</span> npm install @goodganglabs/lipsync-wasm-v2</div>
@@ -750,7 +751,8 @@ <h2 class="section-title">Choose your engine</h2>
       <tr><td>Post-processing</td><td>OneEuroFilter + anatomical constraints</td><td>crisp_mouth + fade + auto-blink</td></tr>
       <tr><td>Idle expressions</td><td>Built-in IdleExpressionGenerator</td><td>Blink injection in post-process</td></tr>
       <tr><td>Voice activity</td><td>Built-in VoiceActivityDetector</td><td>&mdash;</td></tr>
-      <tr><td>Best for</td><td>Full expression control, custom avatars</td><td>Quick integration, lightweight</td></tr>
+      <tr><td>Emotion control</td><td>&mdash;</td><td>5-dim FiLM conditioning (neutral, joy, anger, sadness, surprise)</td></tr>
+      <tr><td>Best for</td><td>Full expression control, custom avatars</td><td>Emotion-aware lip sync, quick integration</td></tr>
     </tbody>
   </table>
 </section>

llms-full.txt

@@ -159,8 +159,9 @@ interface ProcessResult {
 | VRM mode | getVrmFrame() + convert_arkit_to_vrm() for VRM 18-dim | getFrame() only (52-dim ARKit) |
 | Voice activity | Built-in VoiceActivityDetector | Not included |
 | ONNX fallback | Heuristic mode (energy-based) | None (ONNX required) |
+| Emotion control | Not included | 5-dim FiLM conditioning (neutral, joy, anger, sadness, surprise) via setEmotion()/getEmotion() |
 | Body motion | VRMA idle/speaking + VAD auto-switch (LoopPingPong, asymmetric crossfade) | VRMA idle/speaking (LoopPingPong, asymmetric crossfade 0.8s/1.0s) |
-| Best for | Full expression control, custom avatars | Quick integration, lightweight |
+| Best for | Full expression control, custom avatars | Emotion-aware lip sync, quick integration |
 
 ---
 

llms.txt

@@ -14,7 +14,7 @@ AnimaSync extracts emotion from speech and generates lip sync, facial expression
 ## Two Engine Versions
 
 - **V1 (Recommended)**: Phoneme classification, 111-dim ARKit output, built-in VAD
-- **V2 (Lightweight)**: Student distillation model, 52-dim ARKit output, direct prediction
+- **V2 (Lightweight)**: Student distillation model, 52-dim ARKit output, direct prediction, 5-dim emotion control (FiLM conditioning)
 
 ## Quick Start