Chat Interface

Overview

The chat interface lets you ask questions about your infrastructure in plain English. It queries the knowledge graph, assembles context packs, and sends them to a configured LLM for a grounded response.

Endpoints

# WebSocket (streaming responses)
ws://localhost:8080/api/v1/chat

# REST (single response)
POST /api/v1/chat/message
Content-Type: application/json
{"message": "What depends on the production database?"}

How It Works

1. Your question is analyzed to identify relevant resources
2. A context pack is assembled from the knowledge graph (resource details, relationships, documentation)
3. The context pack and your question are sent to the configured LLM
4. The LLM generates a response grounded in your actual infrastructure data

Example Questions

• "What resources are in us-east-1?"
• "Which services depend on the orders database?"
• "Show me all unencrypted EBS volumes"
• "What changed in the last discovery run?"
• "Give me a runbook for the API gateway service"
• "What is our estimated monthly AWS spend?"

Configuration

The chat uses the same LLM configuration as the enrichment service. Configure it in your config file:

llm:
  provider: "openrouter"
  base_url: "https://openrouter.ai/api/v1"
  api_key: "$" + "{OPENROUTER_API_KEY}"
  model: "moonshotai/kimi-k2.5"
  max_tokens: 4096