NeuralMemory for VS Code

A visual brain explorer, inline recall, and memory management extension for NeuralMemory — the reflex-based memory system for AI agents.

Features

Memory Tree View

Browse your brain's neurons directly in the sidebar. Memories are grouped by type (Concepts, Entities, Actions, Time, State, Spatial, Sensory, Intent) with counts and relative timestamps. Click any neuron to instantly recall related memories.

Graph Explorer

Visualize your entire brain as an interactive force-directed graph. Neurons are color-coded by type, synapses show weighted connections. Double-click any node to zoom into its neighborhood.

Drag to pan, scroll to zoom
Click nodes for details and quick recall
Respects VS Code dark/light themes

Eternal Context

Project decisions, instructions, and tech stack are stored directly in the neural graph — fully discoverable via recall and spreading activation. No more JSON sidecar files.

Save project context, decisions, and instructions that persist across sessions
Recap at session start with configurable detail levels (quick / detailed / full)
Topic-based recall for specific areas (e.g., "auth", "database")

Codebase Indexing

Index your codebase into neural memory for code-aware recall. Scans Python files and creates neurons for functions, classes, and imports.

"Where is X implemented?" queries work after indexing
Keyboard shortcut: Ctrl+Shift+M I

External Memory Import

Import memories from other systems into NeuralMemory:

ChromaDB, Mem0, AWF, Cognee, Graphiti, LlamaIndex
Full and incremental sync support

Encode Memories from the Editor

Select any text and encode it as a memory, or use comment triggers (remember:, note:, decision:, todo:) to get inline suggestions via CodeLens.

CodeLens Integration

Functions and classes show memory counts inline. Click to recall related memories or encode new ones. Works across Python, TypeScript, JavaScript, Go, Rust, Java, and C#.

Real-Time Sync

WebSocket connection keeps your tree view, graph, and status bar updated in real time as memories are created or modified from any source (CLI, MCP, other editors).

Commands

Command	Shortcut	Description
Encode Selection as Memory	`Ctrl+Shift+M E`	Encode selected text with optional tags
Encode Text as Memory	—	Type and encode memory content
Recall Memory	`Ctrl+Shift+M R`	Query brain with selectable search depth
Open Graph Explorer	`Ctrl+Shift+M G`	Interactive neuron/synapse visualization
Index Codebase	`Ctrl+Shift+M I`	Scan and index code into neural memory
Recap Session Context	`Ctrl+Shift+M C`	Load saved context at session start
Recap by Topic	—	Search for specific topic in context
Save Eternal Context	—	Save project decisions/instructions
Eternal Context Status	—	View memory counts and session state
Import Memories	—	Import from ChromaDB, Mem0, AWF, etc.
Switch Brain	Click status bar	Switch between local brains
Create Brain	—	Create a new isolated brain
Refresh Memory Tree	Tree header icon	Force refresh from server
Start Server	—	Start local NeuralMemory server
Connect to Server	—	Connect to a remote server

On macOS, use Cmd instead of Ctrl.

Recall Workflow

Trigger recall (Ctrl+Shift+M R) and type your query
Select search depth: Auto, Instant, Context, Habit, or Deep
Choose from matched memories:
- Paste into the active editor
- Copy to clipboard
- View details with confidence score, latency, and matched fiber IDs

Requirements

NeuralMemory v0.9.5+ (pip install neural-memory)
Python 3.11+
A configured brain (nmem init or nmem brain create my-brain && nmem brain use my-brain)

Configuration

Setting	Default	Description
`neuralmemory.pythonPath`	`"python"`	Python interpreter with neural-memory installed
`neuralmemory.autoStart`	`false`	Auto-start the server on activation
`neuralmemory.serverUrl`	`"http://127.0.0.1:8000"`	NeuralMemory server URL
`neuralmemory.graphNodeLimit`	`1000`	Max nodes in graph explorer (50-10000)
`neuralmemory.codeLensEnabled`	`true`	Show CodeLens hints for functions and comments
`neuralmemory.commentTriggers`	`["remember:", "note:", "decision:", "todo:"]`	Comment patterns that trigger encode suggestions

Getting Started

Install the extension
Install NeuralMemory: pip install neural-memory
Run setup: nmem init (creates config, brain, and auto-configures MCP)
Start the server: run NeuralMemory: Start Server from the command palette, or enable neuralmemory.autoStart
Open the NeuralMemory sidebar (brain icon in the activity bar)

Status Bar

The status bar shows your active brain and live statistics:

$(brain) my-brain | N:512 S:1024 F:256

N = Neurons, S = Synapses, F = Fibers
Click to switch brains
Updates every 30 seconds (or instantly via WebSocket)

Support

If you find NeuralMemory useful, consider buying me a coffee:

License

MIT