Skip to main content

Your First Chat

Once you've set up an environment and connected integrations, you can start chatting with AI about your infrastructure.

Prerequisites​

Before your first chat, make sure you have:

  • Created an environment
  • Connected at least one integration (GitHub, GitLab, etc.)
  • Run discovery to populate entities

If you haven't done this yet, follow the First Environment guide.

Starting a Chat​

Navigate to your environment dashboard at app.sixdegree.ai and click the Chat tab. Type your question in the input box and press Enter or click Send.

For terminal users, launch interactive chat mode with:

degree chat

How Chat Works​

When you ask a question, the AI interprets what you're looking for and queries your knowledge graph for relevant entities and relationships. If you've configured MCP tools and an action is needed, the AI can execute those tools. Finally, it generates a natural language response with full context from your infrastructure.

Try These Questions​

For Developers

"Show me the architecture of the auth-service and what it depends on"

"What services will break if I change the payment API?"

"Who maintains the user-service and when was it last deployed?"

For Operators

"What was deployed to production in the last 30 minutes?"

"Which services are using log4j version 2.14?"

"Show me everything that depends on the payment database"

For Product

"When was the new checkout flow deployed to production?"

"What teams will be affected if we deprecate the v1 API?"

"How many deployments did we do this sprint?"

Chat Features​

Contextual Conversations​

The AI maintains context throughout your conversation. You can ask follow-up questions naturally:

You: Show me all Python repositories AI: I found 18 Python repositories in your organization... You: Which ones were updated this week? AI: Of those 18, these 3 were updated in the last 7 days...

Entity References​

Click on any entity name in the AI's response to view its full details in the ontology visualization.

Taking Actions​

If you've configured MCP tools, the AI can execute actions for you:

"Create a GitHub issue in api-service about adding rate limiting"

"Trigger a deployment of the staging environment"

"Check Datadog for errors in the last hour"

Tips for Better Answers​

Be specific rather than vague. Instead of "show me stuff", ask "show me all GitHub repositories that were updated this month".

Use natural language as if talking to a colleague. "What's the status of our production deployment?" works better than trying to construct a query.

Ask for explanations when you need context. "Explain the relationship between api-service and auth-db" helps you understand your architecture.

Troubleshooting​

If the AI can't find anything, check that discovery has run successfully and you're in the correct environment. You can trigger discovery manually from Environment Settings if needed.

If responses seem inaccurate, try being more specific or rephrase your question. Verify that the discovered entities contain the data you're asking about.

If MCP tools aren't working, verify the endpoints are configured correctly and your API tokens and credentials are valid.

Next Steps​

Need Help?​