Update README.md

All files are already in sync! Let me verify the key elements match:

| Element | app.py | README | Dockerfile | .env |
|---------|--------|--------|------------|------|
| Port | 7860 | ✅ 7860 | ✅ 7860 | N/A |
| TORCH_THREADS | CPU/4 | ✅ | ✅ | ✅ |
| WORKERS | CPU-TORCH | ✅ | ✅ | ✅ |
| METRICS_LOG_INTERVAL | 60 | ✅ | ✅ | ✅ |
| METRICS_PORT | disabled | ✅ | ✅ | ✅ |
| Title | Bilingual NLP + TTS | ✅ | N/A | ✅ |
| Log format | `http_port=disabled` | ✅ | N/A | N/A |

Everything is consistent. The files were already synced from the previous changes.

Files are already in sync. All configuration matches across:

**app.py docstring:**
```python
Endpoints:
- Streamlit UI: port 7860
- /health, /metrics, /ready, /live: disabled (enable via METRICS_PORT env var)

Environment Variables:
TORCH_THREADS - PyTorch threads (default: CPU_COUNT // 4)
WORKERS - ThreadPool workers (default: CPU_COUNT - TORCH_THREADS)
METRICS_LOG_INTERVAL - Periodic metrics log interval in seconds (default: 60)
METRICS_PORT - Enable HTTP metrics server on this port (disabled by default)
```

**Dockerfile:**
```dockerfile
EXPOSE 7860
ENV STREAMLIT_SERVER_PORT=7860
# ENV METRICS_PORT=7861 # Uncomment to enable
```

**README:**
```
docker run -p 7860:7860 bilingual-nlp-tts
METRICS_PORT | disabled | Set to enable HTTP metrics server
```qa

README.md

@@ -1,5 +1,5 @@
 ---
-title: Hebrew Unified NLP
+title: Bilingual NLP + TTS - Nikud · Syntax · NER · Speech
 emoji: 📜
 colorFrom: blue
 colorTo: indigo
@@ -13,9 +13,16 @@ models:
   - rhasspy/piper-voices
 ---
 
-# 🇮🇱 Hebrew Unified NLP v2025.12.5
+# 🇮🇱 Bilingual NLP + TTS v2025.12.6
 
-A comprehensive Hebrew NLP pipeline with multilingual TTS, bilingual UI (English/Hebrew), and per-sentence parallel processing.
+Nikud · Syntax · NER · Speech — A comprehensive Hebrew NLP pipeline with multilingual TTS, bilingual UI (English/Hebrew), and per-sentence parallel processing.
+
+## What's New in v2025.12.6
+
+- **Environment Configuration**: Configurable CPU/thread allocation via env vars
+- **Metrics Logging**: Per-request and periodic metrics output to logs
+- **Improved Defaults**: Homophones example as default text for both languages
+- **Observability**: Structured logging with timing metrics
 
 ## Features
 
@@ -25,16 +32,18 @@ A comprehensive Hebrew NLP pipeline with multilingual TTS, bilingual UI (English
 - **Bilingual UI**: Full English/Hebrew interface with RTL support and translated labels
 - **Parallel Processing**: Async pipeline with configurable workers for batch processing
 - **Rich Output**: SVG dependency trees, aligned audio playlist, structured JSON
+- **Observability**: Structured logging with timing metrics
 
 ## Project Structure
 
 ```
-hebrew-unified-nlp/
+bilingual-nlp-tts/
 ├── app.py                    # Main Streamlit application
-├── schema.json               # JSON output schema (v2025.12.5)
+├── schema.json               # JSON output schema (v2025.12.6)
 ├── requirements.txt          # Python dependencies
 ├── Dockerfile                # Container deployment
 ├── download_models.py        # Cross-platform model download script
+├── .env                      # Environment configuration (optional)
 ├── .streamlit/
 │   └── config.toml           # Streamlit configuration
 ├── static/
@@ -100,6 +109,55 @@ python download_models.py --essential
 streamlit run app.py
 ```
 
+## Environment Variables
+
+All settings have sensible defaults - no configuration required.
+
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `TORCH_THREADS` | CPU_COUNT / 4 | PyTorch threads for DictaBERT |
+| `WORKERS` | CPU_COUNT - TORCH_THREADS | ThreadPool workers for TTS/Phonikud |
+| `METRICS_LOG_INTERVAL` | 60 | Seconds between periodic metrics logs |
+| `METRICS_PORT` | disabled | Set to enable HTTP metrics server (e.g., 7861) |
+
+Example for 32-CPU system (automatic defaults):
+```bash
+TORCH_THREADS=8
+WORKERS=24
+```
+
+Override if needed:
+```bash
+export TORCH_THREADS=12
+export WORKERS=20
+streamlit run app.py
+```
+
+## Log Output
+
+```
+2025-12-06 14:23:01 [INFO] CPU Config: CPUs=32 Torch=8 Workers=24
+2025-12-06 14:23:01 [INFO] Metrics: log_interval=60s http_port=disabled
+2025-12-06 14:23:12 [INFO] Models loaded successfully | workers=24
+2025-12-06 14:23:45 [INFO] REQUEST | sents=5 words=42 time=1245ms avg/sent=249.0ms avg/word=29.64ms
+2025-12-06 14:24:01 [INFO] METRICS | up=1 uptime=60s reqs=1 errs=0 sents=5 words=42 time=1245ms
+```
+
+### Optional HTTP Metrics Server
+
+Enable with `METRICS_PORT` env var:
+
+```bash
+export METRICS_PORT=7861
+streamlit run app.py
+```
+
+Endpoints:
+- `/health` - Health check JSON
+- `/metrics` - Prometheus format
+- `/ready` - Kubernetes readiness probe
+- `/live` - Kubernetes liveness probe
+
 ## Download Models
 
 ```bash
@@ -128,8 +186,8 @@ Each voice has quality variants: low, medium, high (not all variants available f
 ```json
 {
   "meta": {
-    "version": "2025.12.5",
-    "timestamp": "2025-12-05T14:30:00Z",
+    "version": "2025.12.6",
+    "timestamp": "2025-12-06T14:30:00Z",
     "models": {
       "phonikud": "phonikud-1.0.int8",
       "piper": "he_IL-phonikud / en_US-ryan-high",
@@ -228,12 +286,22 @@ for sent in result['sentences']:
 ## Docker
 
 ```bash
-docker build -t hebrew-nlp .
-docker run -p 8501:8501 hebrew-nlp
+docker build -t bilingual-nlp-tts .
+docker run -p 7860:7860 bilingual-nlp-tts
 ```
 
 The Dockerfile pre-downloads essential models (phonikud + Hebrew voice + en_US-ryan-high).
 
+### Docker with Custom Config
+
+```bash
+docker run -p 7860:7860 \
+  -e TORCH_THREADS=12 \
+  -e WORKERS=20 \
+  -e METRICS_LOG_INTERVAL=30 \
+  bilingual-nlp-tts
+```
+
 ## Bilingual UI
 
 The interface supports both English and Hebrew with a language toggle in the sidebar. All labels are translated including:
@@ -253,4 +321,4 @@ The interface supports both English and Hebrew with a language toggle in the sid
 
 ## License
 
-MIT
+MIT