75c67150d0
build / build (3.13) (push) Waiting to run
release-please / release-please (push) Waiting to run
release-please / build wheels (macos-aarch64) (push) Blocked by required conditions
release-please / build wheels (macos-x86_64) (push) Blocked by required conditions
release-please / build wheels (windows-x86_64) (push) Blocked by required conditions
release-please / build wheels (linux-aarch64) (push) Blocked by required conditions
release-please / build wheels (linux-x86_64) (push) Blocked by required conditions
release-please / build sdist (push) Blocked by required conditions
release-please / publish release artifacts (push) Blocked by required conditions
memu-cli
🚧 Under heavy construction — memU is undergoing a major rework; commands and flags may change. Expected to stabilize around July 15, 2026.
CLI for memU — personal memory as files: fast retrieval, higher accuracy, lower cost.
This package is a thin launcher. The engine is the Python package memu-py (Python ≥ 3.13); the shim delegates to uvx, pipx run, or an installed python3 -m memu, in that order. Install uv for the smoothest zero-setup experience.
Usage
export OPENAI_API_KEY=sk-...
# diff-sync a workspace folder (chat/ -> memory, agent/ -> skills, other -> index)
npx memu-cli memorize-workspace ./workspace
# single-shot embedding retrieval (LLM-free, fast)
npx memu-cli retrieve-workspace "deploy checklist"
# rebuild the INDEX.md / MEMORY.md / SKILL.md markdown tree
npx memu-cli export
# legacy pair: memorize a single file / LLM-routed retrieval
npx memu-cli memorize notes/meeting.md
npx memu-cli retrieve "What are this user's launch preferences?"
State persists in a local SQLite database (./data/memu.sqlite3 by default), so memorize in one invocation and retrieve in the next. Run npx memu-cli --help or see the main README for all flags and MEMU_* environment variables.
License
Apache-2.0