记忆宫殿 是 AI Skill Hub 本期精选MCP工具之一。综合评分 8.5 分,整体质量较高。我们强烈推荐将其纳入你的 AI 工具库,帮助提升工作效率。
记忆宫殿 是一款遵循 MCP(Model Context Protocol)标准协议的 AI 工具扩展。通过 MCP 协议,它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务,实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用,都可以通过自然语言在 AI 对话中直接触发,极大提升生产效率。
记忆宫殿 是一款遵循 MCP(Model Context Protocol)标准协议的 AI 工具扩展。通过 MCP 协议,它可以让 Claude、Cursor 等主流 AI 客户端直接访问和操作外部工具、数据源和服务,实现 AI 能力的无缝扩展。无论是文件操作、数据库查询还是 API 调用,都可以通过自然语言在 AI 对话中直接触发,极大提升生产效率。
# 方式一:通过 Claude Code CLI 一键安装
claude skill install https://github.com/quangdang46/mempalace_rust
# 方式二:手动配置 claude_desktop_config.json
{
"mcpServers": {
"----": {
"command": "npx",
"args": ["-y", "mempalace_rust"]
}
}
}
# 配置文件位置
# macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
# Windows: %APPDATA%/Claude/claude_desktop_config.json
# 安装后在 Claude 对话中直接使用 # 示例: 用户: 请帮我用 记忆宫殿 执行以下任务... Claude: [自动调用 记忆宫殿 MCP 工具处理请求] # 查看可用工具列表 # 在 Claude 中输入:"列出所有可用的 MCP 工具"
// claude_desktop_config.json 配置示例
{
"mcpServers": {
"____": {
"command": "npx",
"args": ["-y", "mempalace_rust"],
"env": {
// "API_KEY": "your-api-key-here"
}
}
}
}
// 保存后重启 Claude Desktop 生效
<img src="assets/mempalace_logo.png" alt="MemPalace" width="280">
rmcp-macros)rusqlite)install.sh)No vector database server, no API key, no internet after install. The palace ships with an in-process embedvec backend; embeddings run locally via ONNX.
```bash
```bash
curl -fsSL "https://raw.githubusercontent.com/quangdang46/mempalace_rust/main/install.sh?$(date +%s)" | bash
powershell
irm "https://raw.githubusercontent.com/quangdang46/mempalace_rust/main/install.ps1" | iex
bash
cargo install --path . ```
The install.sh script automatically detects your installed AI tools and registers mpr as an MCP server — no manual config editing needed:
```bash curl -fsSL "https://raw.githubusercontent.com/quangdang46/mempalace_rust/main/install.sh?$(date +%s)" | bash
```
Supports: Claude Code, Codex, Cursor, Windsurf, VS Code, Gemini, OpenCode, Amp, Droid
---
mpr init <dir> # guided onboarding + AAAK bootstrap
curl -fsSL "https://raw.githubusercontent.com/quangdang46/mempalace_rust/main/install.sh?$(date +%s)" | bash
Control what gets skipped during mining:
```bash
```
Glob patterns supported. Built-in defaults (.git, node_modules, __pycache__, etc.) still apply unless overridden.
```
Colorized output (green/yellow/red), --no-color for scripts, non-zero exit on failure for CI.
cat ~/.config/mempalace/config.json | jq '.languages = ["pt-BR"]' > /tmp/config.json && mv /tmp/config.json ~/.config/mempalace/config.json
{
"palace_path": "/custom/path/to/palace",
"collection_name": "mpr_drawers",
"people_map": {"Kai": "KAI", "Priya": "PRI"}
}
Generated by mpr init. Maps your people and projects to wings:
{
"default_wing": "wing_general",
"wings": {
"wing_kai": {"type": "person", "keywords": ["kai", "kai's"]},
"wing_driftwood": {"type": "project", "keywords": ["driftwood", "analytics", "saas"]}
}
}
All of these are optional. Defaults are sensible for local single-user use.
| Variable | Default | Purpose |
|---|---|---|
MEMPALACE_PALACE_PATH | ~/.mempalace/palace | Override palace location (also --palace). |
MEMPALACE_NONINTERACTIVE | unset | Skip every prompt with safe defaults — for CI/CD, piped stdin, AI agents. |
MEMPALACE_READONLY | unset | Run MCP server in read-only mode — blocks all mutation tools. |
MEMPALACE_EMBED_MODEL | ONNXMiniLM_L6_V2 | Swap the embedding model (e.g. text-embedding-3-large). |
MEMPALACE_MAX_CHUNKS_PER_FILE | 50000 | Per-file chunk cap. Set 0 to disable; lower it on memory-constrained or older ONNX builds. Overridden by --max-chunks-per-file (upstream #1455). |
MEMPAL_VERBOSE | unset | Verbose tracing on stderr. |
---
mpr mine ~/projects/myapp --exclude "node_modules" --exclude ".log" --exclude "build/*"
mpr search "Clerk decision" --wing driftwood
| Integration | What | Status |
|---|---|---|
| **Hermes** | Memory provider plugin for [Hermes agent framework](https://github.com/NousResearch/hermes-agent) | Planned |
| **OpenClaw** | Skill file for [OpenClaw](https://openclaw.ai) agents (clawhub install mpr) | Planned |
| **AAAK inter-agent** | Use AAAK as token-efficient communication between LLMs | Planned |
| System | LongMemEval R@5 | API Required | Cost |
|---|---|---|---|
| **MemPalace (hybrid)** | **100%** | Optional | Free |
| Supermemory ASMR | ~99% | Yes | — |
| **MemPalace (raw)** | **96.6%** | **None** | **Free** |
| Mastra | 94.87% | Yes (GPT) | API costs |
| Mem0 | ~85% | Yes | $19–249/mo |
| Zep | ~85% | Yes | $25/mo+ |
---
高质量的AI记忆系统实现
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ MIT 协议 — 最宽松的开源协议之一,可自由商用、修改、分发,仅需保留版权声明。
经综合评估,记忆宫殿 在MCP工具赛道中表现稳健,质量优秀。如果你已有明确的使用需求,可以直接上手体验;如果还在评估阶段,建议对比同类工具后再做决策。
| 原始名称 | mempalace_rust |
| 原始描述 | 开源MCP工具:The highest-scoring AI memory system ever benchmarked. Now in Rust with rich fea。⭐8 · Rust |
| Topics | aillmmcpmemorymempalacerust |
| GitHub | https://github.com/quangdang46/mempalace_rust |
| License | MIT |
| 语言 | Rust |
收录时间:2026-05-25 · 更新时间:2026-05-26 · License:MIT · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端