Backend MVP: FastAPI + Claude integration

Core components:
- FastAPI server with health endpoints
- AI client (Anthropic Claude integration)
- Context manager (loads company knowledge)
- Chat API (non-streaming and streaming)
- Requirements and environment setup

Ready to run: python backend/main.py

backend/core/ai_client.py

@@ -0,0 +1,96 @@
+"""
+AI Client - Anthropic Claude Integration
+"""
+
+import anthropic
+import os
+import logging
+from typing import List, Dict, Optional
+
+logger = logging.getLogger(__name__)
+
+class AIClient:
+    """Client for interacting with Anthropic Claude API"""
+    
+    def __init__(self, api_key: str):
+        self.client = anthropic.Anthropic(api_key=api_key)
+        self.model = os.getenv("AI_MODEL", "claude-sonnet-4-5-20250514")
+        self.max_tokens = int(os.getenv("AI_MAX_TOKENS", "4096"))
+        self.temperature = float(os.getenv("AI_TEMPERATURE", "0.7"))
+        logger.info(f"AIClient initialized with model: {self.model}")
+    
+    async def chat(
+        self,
+        messages: List[Dict[str, str]],
+        system_prompt: Optional[str] = None,
+        max_tokens: Optional[int] = None
+    ) -> str:
+        """
+        Send chat messages to Claude and get response
+        
+        Args:
+            messages: List of message dicts with 'role' and 'content'
+            system_prompt: Optional system prompt
+            max_tokens: Optional max tokens override
+            
+        Returns:
+            Response text from Claude
+        """
+        try:
+            kwargs = {
+                "model": self.model,
+                "max_tokens": max_tokens or self.max_tokens,
+                "temperature": self.temperature,
+                "messages": messages
+            }
+            
+            if system_prompt:
+                kwargs["system"] = system_prompt
+            
+            response = self.client.messages.create(**kwargs)
+            
+            # Extract text from response
+            if response.content and len(response.content) > 0:
+                return response.content[0].text
+            
+            return "No response generated"
+            
+        except Exception as e:
+            logger.error(f"Error calling Claude API: {e}")
+            raise
+    
+    async def chat_stream(
+        self,
+        messages: List[Dict[str, str]],
+        system_prompt: Optional[str] = None,
+        max_tokens: Optional[int] = None
+    ):
+        """
+        Stream chat response from Claude
+        
+        Args:
+            messages: List of message dicts
+            system_prompt: Optional system prompt
+            max_tokens: Optional max tokens override
+            
+        Yields:
+            Text chunks from Claude
+        """
+        try:
+            kwargs = {
+                "model": self.model,
+                "max_tokens": max_tokens or self.max_tokens,
+                "temperature": self.temperature,
+                "messages": messages
+            }
+            
+            if system_prompt:
+                kwargs["system"] = system_prompt
+            
+            with self.client.messages.stream(**kwargs) as stream:
+                for text in stream.text_stream:
+                    yield text
+                    
+        except Exception as e:
+            logger.error(f"Error streaming from Claude API: {e}")
+            raise