Back to directory

Compartment

Compartment is encrypted, fully offline persistent memory for AI agents, stored as one vault on the user's own computer that Claude Code, Claude Desktop, Cursor, Codex and other MCP clients can read and write with no API key, account or network connection.

Visit website (opens in a new tab)maxfreedompollard.github.io
Compartment

Compartment stores what an agent learns as individual memories, each recorded with its source and the date it was learned, available across every later session, every project and every agent on the machine. A memory can be given an expiration date and is automatically removed after it, and when a preference changes, a new memory simply replaces the old one; recall is a hybrid vector and keyword search over an in-memory index that answers in about 12 milliseconds.

Everything on disk, including the embedding vectors themselves, is encrypted, and only the user's own passphrase opens the vault; the embedding model ships inside the package so no external service is called. A new vault starts with roughly 6,700 reference facts about hardware, operating systems, ports, encodings and shell tools, which can be excluded from search with one setting.

It installs with pip and a one-command integration step per agent, and its documentation compares its encryption, account requirements and network behavior against several other memory servers as of a stated date. It is written in Python and licensed under Apache 2.0.

More in Agent memory

See all 450
  • A long technical guide to AI agent memory: why stateless LLMs need it, the different memory architectures from vector stores to knowledge graphs, and how Cognee implements one as an open-source memory layer.

  • A local, file-based long-term memory layer for a personal AI that never summarizes conversations, keeping every line with a timestamp and searching by time first and words second, rather than retrieving only similar-seeming chunks.

  • NEOTH

    deepwiki.com

    A local-first personal AI daemon in Rust with five memory tiers, consent-gated tools, a WASM plugin sandbox, multi-provider model routing and signed audit logs, aimed at people who want both a daily assistant and a serious coding helper.