Graph

See how everything connects

Projects grow fast. Docs reference other docs. Tasks depend on code. Names, concepts, and decisions spread across dozens of files. But you only see one file at a time, so connections stay invisible until something breaks.

The Graph applet extracts entities and relationships from your project sources and maps them visually. Pick your docs, code, or stages as inputs, run a build, and get an interactive graph that shows how things relate.

• Connect any source. Pull from docs, code files, stages, and RAG tables. Mix sources in one graph to see cross-resource relationships.
• AI-powered extraction. Go beyond file links. Semantic extraction uses AI to discover entities and relations buried in your content.
• See the shape of your project. Clusters, colors, and shapes make structure visible at a glance. Spot orphaned content, tight dependencies, or unexpected connections.
• Explore interactively. Click any node to see its mentions, neighbors, and metadata. Drag nodes to rearrange. Zoom and pan freely.

Adding a Graph

Open Add Applet and choose Graph in Content.

The creation form lets you configure:

• Name: a label for this graph collection (e.g. “Project Overview”, “Codebase Map”)
• Environment: optional tag for Production, Staging, or Development

After creating the applet, you can create multiple graphs within a single Graph resource, each with its own sources and build configuration.

Selecting Sources

Click Configure Sources in the toolbar to choose what feeds into your graph.

Supported source types:

• Docs: markdown files from your Docs applets
• Code: source code files parsed for functions, classes, and references
• Stages: tasks and cards from your Stage boards
• RAG Tables: indexed content from Local RAG applets

Each source can be toggled on or off. The graph combines all selected sources into a single unified view.

Building the Graph

Once sources are configured, click Build All to extract entities and relationships.

Graph supports three extraction depths:

Depth	Method	What it extracts
Fast	Deterministic	File structure, markdown links, code references, task relations
Semantic	AI-powered	Everything in Fast, plus entities and relations discovered by AI from your content
Full	Premium AI	Everything in Semantic, plus community clustering and hierarchical summaries

Before building, a confirmation dialog shows estimated token usage and cost breakdown per source.

Build results display the number of nodes and edges generated along with the build duration.

Graph Visualization

2D View

The default view renders a force-directed graph using D3.js.

• Nodes represent entities: documents, functions, classes, tags, tasks, stages, and more
• Edges represent relationships: links, references, dependencies, assignments
• Clusters group nodes by source, with a label at each cluster center
• Colors distinguish resource types, with hue shifts per source instance

Node shapes indicate how an entity was extracted:

Shape	Extractor
Circle	Docs
Square	Code
Hexagon	Prose (AI)
Diamond	Stages
Triangle	Tasks

You can drag nodes to rearrange them. Positions are saved automatically.

3D View

An optional 3D mode renders the graph using three.js with orbit controls. Rotate, zoom, and pan to explore the graph from any angle. Labels fade based on camera distance to keep the view readable.

Visual Settings

Customize how the graph looks:

• Color mode: by type (default), by folder, or by link count
• Node size and label font size multipliers
• Show or hide orphans (nodes with no connections)
• Neighbor highlighting on hover
• Source cluster labels toggle

Entity Detail Drawer

Click any node to open the detail drawer on the right side.

The drawer has three tabs:

• Overview: entity type, metadata, and extracted relations
• Mentions: every location where this entity appears across your sources, with file path and line reference
• Neighbors: all directly connected entities (1-hop)

You can pin the drawer open so it stays visible as you explore different nodes.

Data Storage

Graph data is stored locally in your project folder:

rightplace/graph/{resource-name}/{graph-name}/
  graph.db    (SQLite database with nodes, edges, and build metadata)

Each graph maintains its own database. Everything stays local and can be version-controlled.