Memory Engine is permanent memory for AI agents. Store, search, and organize knowledge that persists across conversations.
curl -fsSL https://install.memory.build | shThis installs the me binary to ~/.local/bin. Make sure it's on your PATH.
me loginThis starts an OAuth flow via GitHub -- authorize in your browser and the CLI stores your session.
If your CLI is older than the server (or vice versa), me login will tell you and bail out before sending you to the browser. You can run the same check explicitly:
me versionme memory create "PostgreSQL 18 supports native UUIDv7 generation." \
--tree notes.postgres \
--meta '{"topic": "database"}'# Hybrid search (recommended default: meaning + keywords)
me memory search "UUID generation in Postgres"
# Keyword search
me memory search --fulltext "UUIDv7"
# Pure semantic search (by meaning only)
me memory search --semantic "database-generated identifiers"me memory treeFor a richer, visual experience:
me serveStarts a local web UI on http://127.0.0.1:3000 (or the next free port) with a tree explorer, hybrid / advanced search, rendered Markdown viewer, and a Monaco-based editor for content + metadata. See me serve for details.
Register Memory Engine with your AI coding tools:
me opencode install
me codex install
me gemini installFor Claude Code, install the Memory Engine plugin instead:
claude plugin marketplace add timescale/memory-engine
claude plugin install memory-engine@memory-engineThen start Claude Code, run /plugin, select memory-engine, and configure api_key, server, and tree_prefix.
After installation, your AI agent has access to memory tools -- create, search, get, update, delete, and more.
See MCP Integration for details.
- Core Concepts -- understand memories, tree paths, metadata, search modes
- Access Control -- users, roles, grants, and ownership
- Memory Packs -- install pre-built knowledge collections
- MCP Integration -- how AI agents use Memory Engine
- CLI Reference -- full command reference
- MCP Tool Reference -- full MCP tool reference