Explore Graph

What is the explore graph?

The explore graph is an interactive visualization of your entire codebase structure. It projects your code embeddings into a 2D space using UMAP dimensionality reduction, then overlays the detected interface connections between services.

Each node represents a code entity — a function, class, method, or file. Each edge represents a detected connection between entities, like a REST API call from a frontend to a backend service.

Accessing the graph

Navigate to Explore in your organization's sidebar. The graph loads automatically from your indexed repositories.

Interacting with the graph

• Pan and zoom — scroll to zoom, drag to pan
• Click a node — opens a detail panel showing the file path, symbol name, signature, language, and chunk type
• Click an edge — opens a connection detail panel showing the interface type, match status, confidence score, and both connected nodes
• Fit to View — resets the viewport to show all nodes

Filtering

The sidebar provides filters to focus on specific parts of your codebase:

• Repository — show only nodes from specific repos
• Language — filter by programming language
• Chunk type — filter by entity type (function, class, method, etc.)
• Interface type — filter edges by connection type (REST, Kafka, BullMQ, gRPC, WebSocket, Shared DB)
• Match status — filter edges by confidence (confirmed, inferred, unmatched)

Understanding connections

Edges are color-coded by interface type and labeled with a confidence status:

• Confirmed — both producer and consumer matched with high confidence
• Inferred — connection detected but not fully verified
• Unmatched — producer or consumer detected without a matching counterpart

Repo colors

Each connected repository is assigned a unique color. The legend in the sidebar shows which color maps to which repository, making it easy to see cross-repo connections at a glance.