Knowledge Graph

Not flat text. An interconnected web of knowledge that grows with you.

Nodes

Every piece of knowledge becomes a node in the graph. Each node has:

• Content — the actual knowledge ("prefers TypeScript", "project uses PostgreSQL")
• Type — fact, preference, decision, task, project, skill, person, or event
• Importance — a weight from 0 to 1, reflecting how significant this knowledge is
• Confidence — how certain the brain is about this fact (0-1)
• Timestamps — when it was created, last accessed, last modified

$ brainbase list --limit 5
● a1b2c3d4  preference  "prefers functional components"    2h ago  1.0
● e5f6g7h8  fact        "project uses PostgreSQL"           4h ago  0.9
○ i9j0k1l2  skill       "experienced with TypeScript"       1d ago  0.8

Edges

Nodes are connected by edges — weighted relationships. An edge between "TypeScript" and "React" means the brain knows these are related. Edge types include uses, prefers, contradicts, relates_to, and more.

Edge weights (0-1) strengthen when nodes are used together (Hebbian learning: "fire together, wire together") and weaken over time if unused. Strong edges mean strong associations.

How the graph grows

Not everything gets stored. The Thalamus filters incoming information — noise gets blocked, only important signals pass through. This keeps the graph clean and relevant.

After 1 week, your graph might have 50-100 nodes. After a month, hundreds. After 3 months, a rich web of interconnected knowledge that lets the brain anticipate what you need before you ask.

Explore your graph

$ brainbase search "react"     # Search the graph
$ brainbase list              # Recent nodes
$ brainbase stats             # Graph statistics
$ brainbase dashboard         # 3D visualization