mempalace--mempalace
Local-first AI memory system — 96.6% R@5 trên LongMemEval, verbatim storage (không lossy summary), MCP server 29 tools, ChromaDB/Qdrant/pgvector backends. Zero API calls.
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-14
- Stars
- ★ 13
- Forks
- 2
01 前言 Front Matter
- name
- mempalace--mempalace
- description
- Local-first AI memory system — 96.6% R@5 trên LongMemEval, verbatim storage (không lossy summary), MCP server 29 tools, ChromaDB/Qdrant/pgvector backends. Zero API calls.
- allowed-tools
- Bash, Read, Write
- user-invocable
- true
03 SKILL.md 正文
MemPalace: hệ thống bộ nhớ AI local-first, benchmark cao nhất open-source — lưu verbatim (không tóm tắt), retrieval theo không gian (wing/room/drawer).
Install
uv tool install mempalace
# hoặc
pipx install mempalace
Cách dùng
# Index project hoặc Claude Code conversation archives
mempalace mine ./project/
# Semantic search
mempalace search "database migration strategy"
# Load context cho session mới
mempalace wake-up
Architecture
Palace structure:
Wing — người, project (top-level namespace)
Room — topic trong wing
Drawer — content gốc verbatim
Backends:
ChromaDB (default, local)
SQLite (exact match, correctness verification)
Qdrant (REST external)
pgvector (PostgreSQL)
MCP Server
{
"mcpServers": {
"mempalace": {
"command": "mempalace",
"args": ["mcp"]
}
}
}
29 tools: palace operations, cross-wing navigation, agent diaries, search.
Performance
96.6% R@5 on LongMemEval — zero API calls (embedding only)
98.4% R@5 — hybrid pipeline (keyword + temporal boost)
>99% — với LLM reranking (Claude/Ollama compatible)
Tính năng nổi bật
- Verbatim storage — không lossy extraction, không paraphrase
- Knowledge graph — temporal entity-relationship với validity windows
- Auto-save hooks — periodic save + pre-compression backup cho Claude Code
- Agent diaries — per-agent wings riêng, runtime discovery
- Privacy — không có gì rời máy trừ khi opt-in
So với Memory OS
| MemPalace | Memory OS | |
|---|---|---|
| Storage | Verbatim | Summarized |
| Benchmark | 96.6% LongMemEval | N/A |
| Backends | 4 pluggable | Qdrant only |
| MCP tools | 29 | 16 |
| Layers | 1 (palace metaphor) | 7 |
Source
https://github.com/MemPalace/mempalace · MIT · +446⭐ today
04 同仓库更多 Skills
openai--sentry--sentry
Use when the user asks to inspect Sentry issues or events, summarize recent production errors, or pull basic Sentry health data via the Sentry API; perform read-only queries with the bundled script and require `SENTRY_AUTH_TOKEN`.
ito-basket-compare
Compare Itô prediction-market baskets against a user's knowledge base, portfolio notes, financial context, watchlist, or research thesis. Use for read-only basket comparison and gap analysis without investment advice or live trading.
implementing-network-traffic-baselining
Build network traffic baselines from NetFlow/IPFIX data using Python pandas for statistical analysis, z-score anomaly detection, and hourly/daily traffic pattern profiling
detecting-evasion-techniques-in-endpoint-logs
Detects defense evasion techniques used by adversaries in endpoint logs including log tampering, timestomping, process injection, and security tool disabling. Use when investigating suspicious endpoint behavior, building detection rules for evasion tactics, or conducting threat hunting for stealthy adversary activity. Activates for requests involving evasion detection, defense evasion analysis, log tampering detection, or MITRE ATT&CK TA0005.
implementing-microsegmentation-with-guardicore
Implementing microsegmentation using Akamai Guardicore Segmentation to map application dependencies, create granular network policies, visualize east-west traffic flows, and enforce least-privilege communication between workloads across data centers and cloud.
ramdisk-tmpfs-patterns
暂无简介