A sophisticated multi-agent AI system built on Cloudflare Workers and Containers that orchestrates different specialized AI agents to collaborate on complex research and analysis tasks.

The system consists of a main orchestrator service and three specialized agent containers:

• Purpose: Coordinates agent communication and task distribution
• Technology: Cloudflare Workers, Hono framework, TypeScript
• Responsibilities: Request routing, response aggregation, error handling

• Capabilities:
  • Text summarization using BART transformer
  • Sentiment analysis with RoBERTa
  • Named entity recognition with spaCy
  • Language detection
• Technology: FastAPI, transformers, spaCy, NLTK, TextBlob

• Capabilities:
  • Web scraping and content extraction
  • Search result simulation
  • Content parsing with Readability
  • URL validation and metadata extraction
• Technology: Node.js, Hono, Cheerio, Puppeteer, Readability

• Capabilities:
  • Statistical analysis (mean, median, std dev)
  • Trend detection and correlation analysis
  • Data insights generation
  • Performance metrics calculation
• Technology: Go, Gin framework, Gonum statistics

Submit a research query and get:

• Web search results with extracted content
• Sentiment analysis of findings
• Statistical analysis of data patterns
• Comprehensive insights and trends

• Extract and analyze text from multiple sources
• Generate summaries and key insights
• Identify correlations and patterns
• Produce actionable intelligence

• Process structured and unstructured data
• Generate statistical reports
• Detect trends and anomalies
• Provide automated insights

System information and available endpoints

System health check with agent status

List all available agents and their capabilities

Submit a comprehensive research request

Request Body:

{
  "query": "artificial intelligence in healthcare",
  "focus_areas": ["machine learning", "diagnostics"],
  "max_sources": 5,
  "include_sentiment": true,
  "include_analytics": true
}

Response:

{
  "success": true,
  "data": {
    "query": "artificial intelligence in healthcare",
    "summary": "AI in healthcare is transforming...",
    "sources": [...],
    "sentiment_analysis": {
      "score": 0.8,
      "label": "positive",
      "confidence": 0.95
    },
    "key_entities": [...],
    "analytics": {
      "statistics": {...},
      "trends": [...],
      "insights": [...]
    },
    "generated_at": "2025-06-24T20:15:30Z",
    "processing_time_ms": 3450
  }
}

Test NLP agent directly

Test Web agent directly

Test Analytics agent directly

• Node.js 20+
• pnpm
• Docker (for container development)
• Cloudflare Workers CLI (wrangler)

# Install dependencies
pnpm install

# Generate Cloudflare types
pnpm run cf-typegen

# Start development server
pnpm run dev

# Run all tests
pnpm test

# Run linting and type checking
pnpm run check

Each agent can be developed and tested independently:

# Test NLP Agent locally
cd containers/nlp
pip install -r requirements.txt
python main.py

# Test Web Agent locally
cd containers/web
npm install
npm start

# Test Analytics Agent locally
cd containers/analytics
go mod download
go run .

# Deploy to Cloudflare
pnpm run deploy

• Each container runs on port 8080
• NLP Agent requires ~2GB RAM for transformer models
• Web Agent needs internet access for scraping
• Analytics Agent is lightweight and CPU-focused

• Simple NLP tasks: 100-500ms
• Web scraping (3-5 URLs): 2-5 seconds
• Statistical analysis: 50-200ms
• Full research pipeline: 3-10 seconds

• Horizontal scaling: Multiple container instances
• Auto-scaling: Based on request volume
• Resource isolation: Each agent in separate containers
• Fault tolerance: Individual agent failures don't crash system

• Container isolation: Each agent runs in isolated environment
• No persistent storage: Stateless request processing
• Input validation: Comprehensive request validation
• Error sanitization: No sensitive data in error messages

curl -X POST {HOST}/research \
  -H "Content-Type: application/json" \
  -d '{
    "query": "electric vehicle market trends 2024",
    "max_sources": 10,
    "include_sentiment": true,
    "include_analytics": true
  }'

curl -X POST {HOST}/research \
  -H "Content-Type: application/json" \
  -d '{
    "query": "kubernetes security best practices",
    "focus_areas": ["container security", "network policies"],
    "max_sources": 7,
    "include_sentiment": false,
    "include_analytics": true
  }'

curl -X POST {HOST}/research \
  -H "Content-Type: application/json" \
  -d '{
    "query": "renewable energy adoption rates",
    "max_sources": 15,
    "include_sentiment": true,
    "include_analytics": true
  }'

• Built-in health checks for all agents
• Performance metrics in responses
• Error tracking with detailed error messages
• Request tracing through the orchestration pipeline

• Agent Plugin System: Dynamic agent registration
• Caching Layer: Redis-based response caching
• Streaming Responses: Real-time result streaming
• Advanced ML Models: Integration with GPT/Claude APIs
• Visual Analytics: Chart and graph generation
• Multi-language Support: International content analysis

This is a proof-of-concept demonstrating Cloudflare Containers for AI agent orchestration. The system showcases:

• Microservice architecture with language-specific agents
• Container orchestration for specialized AI workloads
• Real-time collaboration between different AI capabilities
• Scalable and fault-tolerant design patterns

Perfect for research institutions, content intelligence platforms, and organizations needing automated analysis workflows.