代码记忆宫殿

Print a scheduler snippet (does NOT install — owner action required)

mempalace-code backup schedule --freq daily # daily at 03:00 mempalace-code backup schedule --freq weekly # weekly on Sunday at 03:00 mempalace-code backup schedule --freq hourly # every hour

Setup

mempalace-code init <dir> # initialize rooms mempalace-code init <dir> --detect-entities # optional prose entity bootstrap

Quick Start

```bash uv tool install mempalace-code # recommended (fast, Rust-based)

Optional Entity Detection

mempalace-code init <dir> is config-first by default: it detects rooms from the directory structure and does not scan file contents for names. Add --detect-entities only when the directory contains prose where people or project names matter, such as meeting notes, client notes, personal notes, or conversation exports:

mempalace-code init ~/notes --detect-entities        # prompts to confirm detected people/projects
mempalace-code init ~/notes --detect-entities --yes  # auto-accept entity confirmation (no room prompts)

The detector is a lightweight bootstrap step, not the main miner. It samples up to 10 readable files, prefers prose files (.md, .txt, .rst, .csv), reads the first 5 KB of each sampled file, and looks for heuristic signals such as Alice said, thanks Bob, Apollo repo, deploy Apollo, or import Apollo. Confirmed results are written to <dir>/entities.json:

{
  "people": ["Alice", "Bob"],
  "projects": ["Apollo"]
}

Use it for human/project context. Leave it off for normal code repos unless their docs contain the entities you want captured. Full-repo scanning would be slower and noisier: class names, packages, examples, and variables often look like people or products to a heuristic pass. Code structure, symbols, languages, and architecture relationships are handled by mempalace-code mine, not by entity detection.

Environment variable (bytes or human suffix)

export MEMPALACE_WATCH_DISK_MIN_FREE_BYTES=2GiB # watcher-specific floor export MEMPALACE_DISK_MIN_FREE_BYTES=1GiB # global floor (watcher + backup)

~/.mempalace/config.json

{ "disk_min_free_bytes": 1073741824, // 1 GiB global default "watch_disk_min_free_bytes": 2147483648, // 2 GiB for watcher specifically "backup_disk_min_free_bytes": 2147483648 // 2 GiB for backups specifically } ```

MCP read/search behavior is not affected by a paused watcher — agents can still search the palace while the daemon waits for disk space.

---

Check right now regardless of the interval setting

mempalace-code version-check --check-now ```

How it works:

• On the first interactive command after a fresh install, the CLI prompts once: "Enable periodic new-version checks?" — answering n records the opt-out permanently. Non-interactive (piped, CI, non-TTY) invocations never prompt.
• When opted in, a background check runs at most once per interval (default: 168 hours / 1 week). Any update hint appears on stderr only — stdout remains machine-parseable.
• Explicit --check-now ignores the interval, contacts PyPI, and prints current/latest/error to stdout.
• Only https://pypi.org/pypi/mempalace-code/json is contacted. No telemetry, no user IDs, no installed-package inventory.

Environment overrides:

Setting MEMPALACE_VERSION_CHECK=0 in a CI pipeline guarantees no network calls regardless of any saved preference.

---

Opt in to periodic checks (contacts PyPI for package metadata only)

mempalace-code version-check --enable

This Fork vs Upstream

This is a code-first fork of milla-jovovich/mempalace. We inherited the good parts — the palace metaphor, the MCP integration, the LongMemEval harness — and rebuilt what was broken. Every claim here is backed by code, tests, and documented benchmarks.

Full audit: docs/UPSTREAM_HARDENING.md.

---

Token savings vs grep + read ([full methodology](docs/BENCH_TOKEN_DELTA.md))

Token savings scale with project size — grep noise grows linearly (more files contain the keyword), while mempalace-code search stays constant (top-5 semantically relevant chunks regardless of corpus size). These numbers are from a 19k-chunk project; larger codebases would push the ratios higher.