Speeding Up the Agent › Putting It All Together

import os
from anthropic import Anthropic
import hashlib
import time

class FastAgent:
    """An optimized agent that prioritizes speed."""
    
    def __init__(self):
        self.client = Anthropic(api_key=os.getenv("ANTHROPIC_API_KEY"))
        self.cache = {}
        self.system_prompt = """You are a helpful assistant. Provide concise, 
        direct answers. Use 1-2 sentences for simple questions."""
    
    def _hash_query(self, query):
        """Create cache key."""
        return hashlib.md5(query.encode()).hexdigest()
    
    def _is_simple_query(self, query):
        """Determine if query is simple enough for fast model."""
        simple_patterns = ["what is", "who is", "when is", "where is"]
        return any(pattern in query.lower() for pattern in simple_patterns)
    
    def respond(self, query):
        """Get fast response using multiple optimization techniques."""
        # 1. Check cache first
        cache_key = self._hash_query(query)
        if cache_key in self.cache:
            return self.cache[cache_key], "cache"
        
        # 2. Choose appropriate model and token limit
        if self._is_simple_query(query):
            max_tokens = 150
            model = "claude-sonnet-4-5"  # Still fast for simple queries
        else:
            max_tokens = 500
            model = "claude-sonnet-4-5"
        
        # 3. Make the call with optimizations
        start = time.time()
        response = self.client.messages.create(
            model=model,
            max_tokens=max_tokens,
            system=self.system_prompt,
            messages=[{"role": "user", "content": query}]
        )
        elapsed = time.time() - start
        
        result = response.content[0].text
        
        # 4. Cache the result
        self.cache[cache_key] = result
        
        return result, f"model ({elapsed:.2f}s)"

## Test the fast agent
agent = FastAgent()

queries = [
    "What is Python?",
    "What is Python?",  # Should hit cache
    "Explain object-oriented programming",
]

for query in queries:
    result, source = agent.respond(query)
    print(f"Q: {query}")
    print(f"A: {result}")
    print(f"Source: {source}\n")