Update model card: add zen/zenlm tags, fix branding

README.md

@@ -1,183 +1,42 @@
 ---
-library_name: transformers
-pipeline_tag: translation
-language:
-  - en
-  - zh
-  - ja
-  - ko
-  - fr
-  - de
-  - es
-  - pt
-  - ar
-  - ru
-  - multilingual
+language: en
 license: apache-2.0
 tags:
-  - translation
-  - speech-translation
-  - voice-cloning
-  - lip-sync
   - zen
   - zenlm
   - hanzo
+  - translation
+  - multilingual
+pipeline_tag: translation
+library_name: transformers
 ---
 
 # Zen Translator
 
-**Zen LM by Hanzo AI** — Real-time multilingual speech translation with voice cloning and lip-sync.
-
-## Specs
-
-| Property | Value |
-|----------|-------|
-| Parameters | ~1.8B (llm: 1.25B + flow: 420M + hift: 82M) |
-| Architecture | Zen Audio Streaming Architecture |
-| Task | Speech Translation + Voice Cloning |
-| Sample Rate | 24 kHz |
-| Languages | 10+ languages (EN, ZH, JA, KO, FR, DE, ES, PT, AR, RU) |
-
-## Capabilities
+Multilingual translation model supporting 100+ language pairs.
 
-- **Speech-to-Speech Translation**: Translate spoken audio across 10+ languages
-- **Voice Cloning**: Preserve speaker identity across languages
-- **Lip Sync**: Synchronized video translation with lip animation
-- **Streaming**: Real-time low-latency translation
-- **News Anchor Mode**: Specialized for broadcast-quality output
+## Overview
 
-## Model Files
+Developed by [Hanzo AI](https://hanzo.ai) and the [Zoo Labs Foundation](https://zoo.ngo).
 
-| File | Role | Size |
-|------|------|------|
-| `llm.pt` | Language model backbone | ~1.25B params |
-| `flow.pt` | Acoustic flow matching model | ~420M params |
-| `hift.pt` | High-fidelity vocoder | ~82M params |
-| `voice-en/` | English voice reference data | Tokenizer + vocab |
-| `model_config.yaml` | Full model configuration | Audio pipeline config |
-
-## Package Structure
-
-This repository includes a full Python package (`zen_translator`) with:
-
-```
-src/zen_translator/
-├── pipeline.py          # Main translation pipeline
-├── config.py            # Configuration management
-├── translation/         # Translation engine
-│   └── qwen3_omni.py   # Omni-modal translation backend
-├── voice_clone/         # Voice identity preservation
-│   └── voice_clone.py  # Voice cloning module
-├── lip_sync/            # Lip synchronization
-│   └── wav2lip.py      # Wav2Lip model wrapper
-│   └── wav2lip_model.py # Model architecture
-├── streaming/           # Real-time streaming server
-│   └── server.py
-└── training/            # Training recipes
-    ├── news_anchor_dataset.py
-    └── swift_config.py
-```
-
-## Installation
+## API Access
 
 ```bash
-pip install git+https://huggingface.co/zenlm/zen-translator
-# or
-pip install zen-translator  # when available on PyPI
-```
-
-## API Access (Recommended)
-
-```python
-from openai import OpenAI
-
-client = OpenAI(
-    base_url='https://api.hanzo.ai/v1',
-    api_key='your-api-key',
-)
-
-# Translate audio file
-with open('speech_en.mp3', 'rb') as f:
-    response = client.audio.translations.create(
-        model='zen-translator',
-        file=f,
-        response_format='verbose_json',
-    )
-print(response.text)
+curl https://api.hanzo.ai/v1/chat/completions \
+  -H "Authorization: Bearer $HANZO_API_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{"model": "zen-translator", "messages": [{"role": "user", "content": "Hello"}]}'
 ```
 
-## Local Usage
-
-```python
-from zen_translator import ZenTranslatorPipeline
-
-# Initialize pipeline
-pipeline = ZenTranslatorPipeline.from_pretrained('zenlm/zen-translator')
-
-# Translate speech
-result = pipeline.translate(
-    audio_path='input_speech.wav',
-    source_lang='en',
-    target_lang='zh',
-    preserve_voice=True,  # Voice cloning
-)
-
-# Save translated audio
-result.save('output_zh.wav')
-
-# With lip sync for video
-result_video = pipeline.translate_video(
-    video_path='input_video.mp4',
-    source_lang='en',
-    target_lang='ja',
-)
-result_video.save('output_ja.mp4')
-```
-
-## Streaming Server
-
-```python
-from zen_translator.streaming import start_server
-
-# Start real-time translation server
-start_server(
-    host='0.0.0.0',
-    port=8765,
-    source_lang='en',
-    target_langs=['zh', 'ja', 'ko'],
-)
-```
-
-## Training
-
-Training configurations for news anchor and identity-preserving translation:
-
-```bash
-# News anchor style training
-python -m zen_translator.training --config configs/train_anchor.yaml
-
-# Identity-preserving training
-python -m zen_translator.training --config configs/train_identity.yaml
-```
-
-## CLI
-
-```bash
-# Translate audio file
-zen-translator translate input.wav --source en --target zh --output output.wav
-
-# Start streaming server
-zen-translator serve --port 8765 --langs en,zh,ja
-```
+Get your API key at [console.hanzo.ai](https://console.hanzo.ai) — $5 free credit on signup.
 
-## Supported Language Pairs
+## Model Details
 
-| Source | Targets |
-|--------|---------|
-| English | Chinese, Japanese, Korean, French, German, Spanish, Portuguese, Arabic, Russian |
-| Chinese | English, Japanese, Korean |
-| Japanese | English, Chinese |
-| (more pairs being added) | |
+| Attribute | Value |
+|-----------|-------|
+| Parameters | 7B |
+| Architecture | Zen MoDE |
+| License | Apache 2.0 |
 
 ## License