Knowledge Graph

One of the most powerful features of l0-memory is its ability to organize information as a Knowledge Graph. Instead of flat notes, your assistant can build a web of interconnected concepts.

Typed, Directional Edges

A link in l0-memory is defined as a triple: (from, to, relationship).

• Directional: Links have a clear "from" and "to".
• Typed: Every link has a rel (relationship) string, such as depends_on, implements, part_of, or related_to.
• Unique: The same triple can only exist once.

Cross-Scope Linking

Links can bridge different scopes. You can link a repo: specific memory to a global user memory.

# Example: Linking a project component to a global technology note
ltm link repo:my-app:auth depends_on user:oauth2-standards

Graph Traversal

AI assistants can "walk" the graph using the memory_traverse tool. This allows them to discover related context without knowing the specific keys in advance.

BFS Algorithm

memory_traverse performs a Breadth-First Search (BFS) starting from a root node.

• Depth: Control how many levels deep the traversal should go (default is 1).
• Direction: Follow links out (outgoing), in (incoming), or both.
• Filtering: Optionally restrict the traversal to specific relationship types.

Visualizing the Graph

The VS Code extension includes a dedicated Knowledge Graph viewer.

• D3.js Powered: Renders a force-directed graph of your memories.
• Interactive: Drag nodes, zoom in/out, and click to open specific memories.
• Scope Coloring: Nodes are automatically colored based on their scope.
• Freshness Indicators: Pinned nodes and the current root node are visually highlighted.

Cascading Deletion

l0-memory maintains referential integrity. When a memory is deleted, all links incident to it (both incoming and outgoing) are automatically removed, preventing "dangling" edges in your graph.