Observee provides an interactive knowledge graph visualization to understand the relationships between your MCP servers, tools, and their interactions. You can access the knowledge graph from your dashboard. The graph displays the flow of data through your MCP ecosystem, showing how servers host tools and how those tools process inputs to generate responses. Screenshot2025 06 26at9 40 01PM Pn ## Understanding the Graph The knowledge graph uses different node types to represent your MCP ecosystem:
  • Server nodes (blue): Your MCP servers
  • Tool nodes (green): Individual tools hosted by servers
  • Input nodes (yellow): Tool inputs/requests
  • Response nodes (red): Tool outputs, colored by status
Click on any node to see detailed information including full input/output text, execution duration, and timestamps. Failed tool calls will be highlighted in red with error details visible in the response nodes. ## Session Graph The session graph provides a chronological view of interactions within a specific session, showing the temporal sequence of tool and prompt calls. Screenshot2025 06 26at9 29 56PM Pn Session graphs display interactions as a timeline, with edges representing the order of execution. Each node represents a single interaction (tool call or prompt) with sequence numbers indicating the order of execution. All inputs and outputs are displayed as JSON strings with full context preserved for analysis.