Skip to content
| Marketplace
Sign in
Visual Studio Code>AI>HackLM MemoryNew to Visual Studio Code? Get it now.
HackLM Memory

HackLM Memory

HackLM

| (1) | Free
Persistent memory layer for VS Code Copilot — learn from interactions, retain insights, improve over time
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Copied to clipboard
More Info

HackLM Memory

HackLM Memory

A long-term memory layer for VS Code Copilot.
Learns from your conversations. Stays local. Gets smarter over time.


What It Does

HackLM Memory gives Copilot a persistent memory across sessions. It stores decisions, preferences, quirks, and security rules in plain Markdown files inside your project.

  • Learns passively — picks up preferences and rules from natural conversation
  • Never stores duplicates — deduplication runs on every write
  • Self-cleaning — periodic cleanup merges similar entries and prunes stale ones
  • Workspace-scoped — each project has its own .memory/ folder
  • Fully local — no cloud sync, no telemetry

Getting Started

  1. Install the extension
  2. Open a workspace — memory files are created automatically in .memory/
  3. Start chatting with Copilot — it will read and write memories automatically
  4. Click Memory in the status bar to open the control panel

Memory Categories

Category What It Stores
Instruction How Copilot should behave in this project
Quirk Non-obvious project-specific knowledge
Preference Style, tone, and design choices
Decision Architectural commitments
Security Rules that must never be broken

Commands

Command Description
HackLM Memory: Open Control Panel Browse all actions in one place
HackLM Memory: List Memories View all stored entries
HackLM Memory: Delete Memory Remove a specific entry
HackLM Memory: Run Cleanup Merge duplicates and prune stale entries
HackLM Memory: Review Session Find decisions not yet captured
HackLM Memory: Open Memory Folder Browse .memory/ files directly

Settings

Setting Default Description
hacklm-memory.lmFamily gpt-5-mini Model family for LM operations
hacklm-memory.autoApproveStore false Skip confirmation when saving memories
hacklm-memory.manageInstructionFile true Manage .github/copilot-instructions.md automatically
hacklm-memory.autoCleanupFrequency 10 Run cleanup automatically every N store operations
hacklm-memory.categoryLimit.* 20 Max entries per category

Privacy

All memory is stored locally in your workspace .memory/ folder. Nothing leaves your machine.
Add .memory/ to .gitignore if any contents are sensitive.


Requirements

  • VS Code 1.99 or later
  • GitHub Copilot (for LM tool integration)

License

GPL-3.0

  • Contact us
  • Jobs
  • Privacy
  • Manage cookies
  • Terms of use
  • Trademarks
© 2026 Microsoft