> For the complete documentation index, see llms.txt.
Skip to main content

Check out Port for yourself ➜ 

Conversational insights

Query and analyze your context lake with AI. Ask questions in natural language, get answers grounded in your catalog data, and turn insights into dashboards or reports — without switching tools or writing queries by hand.

Why it matters

Engineering data lives across dashboards, scorecards, incidents, and ownership records. Extracting answers usually means manual investigation or expertise in multiple tools. Conversational access lets anyone explore delivery trends, investigate incidents, or prepare leadership reports from the same unified context.

How Port helps

Port connects AI to your software catalog & context lake in three ways:

  • AI agents in Port analyze metrics, explain trends, and recommend next steps using data and MCP connectors from GitHub, Jira, PagerDuty, and more.
  • MCP outside Port lets you query the same context from Claude, Cursor, or any MCP-compatible client — ideal for ad-hoc questions during reviews or investigations.
  • Dashboards and reports combine built-in Port widgets, custom widgets, and AI-generated outputs through tools like Claude or Lovable.

Across all three, you can explore service ownership, DORA and delivery metrics, scorecard compliance, incident context, and cross-team comparisons from one connected data layer.

Choose how you interact

AI agents in Port

Use Port's AI agents and custom skills to analyze engineering intelligence data inside the platform. Agents work directly on your catalog and can pull live context from connected tools.

Example uses:

  • Delivery analysis — "Which teams had the biggest increase in PR cycle time this quarter, and what's driving it?"
  • Scorecard remediation — "What do we need to fix to reach Silver on our DORA scorecard?"
  • Reliability review — "What are our top services by incident frequency, and are they improving?"
Implement this use case

Follow the recommended guides below to implement this use case.