🍳 Cookbook: Agentic Translation Engine¶

Build an intelligent translation service that goes beyond simple text—handling PDFs, images (via OCR), and multi-language documents with AI-powered quality checking.

🎯 What You'll Build¶

A service that: 1. Translates text between 20+ languages 2. Processes files (PDF, images, Markdown) 3. Uses AI to validate translation quality

🏗️ Phase 1: Clone & Setup¶

1. Get the Code¶

git clone https://github.com/traylinx/agentic-translation-engines.git
cd agentic-translation-engines

2. Install Dependencies¶

pip install -r requirements.txt

3. Configure Environment¶

cp .env.sample .env
nano .env

Key Variables:

# LLM Configuration
LLM_MODEL=gpt-4o
LLM_API_KEY=your-api-key
LLM_BASE_URL=https://switchai.traylinx.com/v1

# File Engine (Required for PDF/OCR)
FILE_ENGINE_BASE_URL=https://api.traylinx.com/file-engine

🚀 Phase 2: Run the Service¶

uvicorn api:app --reload
# Server running at http://localhost:8000

🧪 Phase 3: Basic Text Translation¶

Simple Translation Request¶

curl -X POST http://localhost:8000/v1/translate \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Hello, how are you today?",
    "source_language": "en",
    "target_language": "es"
  }'

Response:

{
  "translated_text": "Hola, ¿cómo estás hoy?",
  "source_language": "en",
  "target_language": "es",
  "quality_score": 0.95
}

📄 Phase 4: Document Translation¶

Translate a PDF¶

curl -X POST http://localhost:8000/v1/translate/document \
  -H "Authorization: Bearer your-token" \
  -F "file=@contract.pdf" \
  -F "target_language=de"

What Happens: 1. File Engine extracts text (uses OCR if needed) 2. Translator Agent translates content 3. Quality Checker validates accuracy 4. Returns translated document

🔄 Phase 5: The Agentic Workflow¶

The translation engine uses a multi-agent orchestration pattern:

          Document
             │
             ▼
     ┌───────────────┐
     │ File Engine   │ ← Extract text, handle OCR
     └───────┬───────┘
             │
             ▼
     ┌───────────────┐
     │  Translator   │ ← LLM-powered translation
     └───────┬───────┘
             │
             ▼
     ┌───────────────┐
     │Quality Checker│ ← Validates accuracy, suggests fixes
     └───────┬───────┘
             │
     ┌───────┴───────┐
     │  Pass?        │
     └───────┬───────┘
        No   │   Yes
             ▼
       [Retry Loop] → Final Output

The Orchestrator manages retries if the Quality Checker finds issues.

🌍 Supported Languages¶

Language	Code	Language	Code
English	`en`	Spanish	`es`
French	`fr`	German	`de`
Italian	`it`	Portuguese	`pt`
Japanese	`ja`	Korean	`ko`
Chinese (Simplified)	`zh`	Arabic	`ar`
Russian	`ru`	Hindi	`hi`

And many more!

🔐 Phase 6: Enable A2A Authentication¶

Let other agents use your translation service.

Add to .env¶

TRAYLINX_CLIENT_ID=ag-xxx
TRAYLINX_CLIENT_SECRET=ts-xxx

Call as an Agent¶

curl -X POST http://localhost:8000/v1/translate \
  -H "X-Agent-Secret-Token: your-token" \
  -H "X-Agent-User-Id: your-agent-id" \
  -H "Content-Type: application/json" \
  -d '{"text": "Bonjour", "target_language": "en"}'

📚 Next Steps¶

Integrate batch translation for large document sets
Add custom glossaries for domain-specific terms
Deploy to Kubernetes for production scale