MaxFreedomPollard/Compartment

Encrypted, fully offline agentic memory. One click install, GUI w/ memory map, all OS and agents. Superior memory creation, storage and retrieval.

What it solves

Compartment provides a secure, fully offline, and encrypted persistent memory system for AI agents. It solves the problem of agents losing context between sessions or across different projects, while ensuring that sensitive data never leaves the user's local machine and is not subject to telemetry or account requirements.

How it works

Compartment acts as a local vault that stores memories as single claims, each associated with a source and a date. It uses a hybrid search approach combining a local vector index (using a bundled ONNX embedding model) and a keyword index to retrieve relevant information in milliseconds.

Key technical mechanisms include:

  • Encryption: All data on disk, including embedding vectors, is encrypted with a user passphrase.
  • Memory Logic: It enforces a "one claim per memory" rule (max 200 characters) to ensure granularity. It distinguishes between "facts" (which accumulate) and "opinions" (which are updated/superseded).
  • Ranking: A custom mathematical ranking system combines semantic and literal evidence, weighted by importance tiers (e.g., decisions, personal facts) and recency.
  • Integration: It integrates with MCP (Model Context Protocol) clients like Claude Code, Cursor, and VS Code, and can use hooks to automatically capture memories from specific agents.
  • Dynamic Tagging: A background process updates tags based on embedding similarity to keep memories discoverable as project contexts evolve.

Who it’s for

It is designed for users of AI agents (such as Claude Code, Cursor, or Hermes Agent) who require absolute privacy, offline capability, and a way for their agents to maintain a long-term, cross-session memory of preferences, decisions, and technical facts.

Highlights

  • Zero-Trust Privacy: No API keys, no accounts, no network access, and no telemetry.
  • Hybrid Search: Combines vector and keyword search to ensure precise retrieval of specific identifiers (like commit SHAs) and semantic concepts.
  • Local Embeddings: Includes a bundled 384-dimension int8 ONNX model for completely offline processing.
  • Encrypted Vault: Single-file encrypted storage that protects both text and vectors.
  • Management Tools: Includes a system tray app for vault control and a browser-based dashboard for visualizing the memory graph and growth.

Related

  • Project
    doobidoo/mcp-memory-service

    A self‑hosted memory backend for LLM agents that provides a REST API, CLI, OAuth, and a web dashboard. It stores facts, decisions, and causal knowledge graphs locally (SQLite + ONNX embeddings) or in hybrid/cloud mode, offering fast (≈5 ms) semantic search, tag‑scoped retrieval, and optional LLM‑generated summaries. Works with LangGraph, CrewAI, AutoGen, Claude Desktop, ChatGPT, and many other tools.

  • Project
    breferrari/obsidian-mind

    A template vault for Obsidian that adds persistent, searchable memory to AI coding agents (Claude Code, Codex CLI, Gemini CLI) via hooks, slash commands, and optional local semantic search (QMD).

  • Project
    mcncarl/agent-memory-vault

    Agent Memory Vault is a Python toolkit that provides a local, Git‑backed Markdown vault with SQLite (and optional vector) search, enabling Claude Code, Codex and similar AI agents to share durable, auditable long‑term memory across sessions.

  • Project
    Lyellr88/marm-memory

    marm-memory is a local‑first memory server for AI agents that use the Model Context Protocol (MCP). It stores chats, notes, code‑graph data and extracted concepts in SQLite, exposing them via HTTP or STDIO so agents can retrieve relevant context instead of starting from scratch. The package includes a web console, Docker image, and CLI tools for starting, managing, and indexing repositories. Benchmarks show low‑latency retrieval (≈30 ms for 4 k memories) and hybrid FTS‑semantic recall that scales to 10 k entries. Ideal for developers building coding assistants or multi‑agent swarms who want a self‑hosted, searchable memory layer.

  • Project
    rahilp/second-brain-cloudflare

    Second Brain is a Cloudflare‑hosted personal + team memory system that stores notes, decisions, and project facts and makes them searchable by meaning. It integrates with Claude, ChatGPT, Cursor, Obsidian, Notion, calendars, etc., and provides a secure, self‑hosted backend (D1, Vectorize, Workers AI) plus a “Prompt Capsule” feature for deterministic LLM context.