Brain MCP Dcd6b298

MCP运维云端

提供方 yonidavidson

运维云端 未标注 无需鉴权 社区 / 第三方 Claude DesktopClaude

Brain MCP 是一个基于 DuckDB 的智能对话记忆存储服务器,支持本地文件、S3 和 Google Cloud Storage 等多种存储后端,提供对话上下文存储、检索和长期记忆整合功能。

一句话选型结论

按资料字段整理

Brain MCP Dcd6b298 属于运维云端类 MCP 服务器,传输方式未标注,无需鉴权,可直接接入 Claude Desktop、Claude,没有现成安装命令,需按官方文档接入。接入前先确认权限范围;涉及写入、删除或付款的操作,建议先在测试环境验证。

分类运维云端 传输方式未标注 鉴权无需鉴权 部署方式未标注
GitHub 星标 0 仓库热度,无公开仓库时显示未标注
工具数 未标注 暴露给模型的工具数量
资源 / 提示词 0 / 0 资源与提示词模板数量
维护方 社区 / 第三方 核验等级 仅资料核对

工具与能力

运维云端
数量来自来源目录的标注,具体工具清单以官方文档为准
工具数量未标注暴露给模型可调用的工具个数
资源数量0可读取的资源条目
提示词数量0内置提示词模板数量

能力标签

支持客户端

适用场景

智能存储TypeScript本地部署对话管理记忆整合DuckDB多后端支持

安装与配置

未标注 无需鉴权
上游未给出传输方式
  1. 1
    准备运行环境

    按官方文档确认运行环境(常见为 Node.js 或 Python),装好后再执行安装命令。

  2. 2
    安装服务

    复制上面的安装命令执行;命令里带占位符的位置,换成自己的密钥或路径。

  3. 3
    写入客户端配置

    在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,让新配置生效。

  4. 4
    验证是否可用

    连接成功后随便调一个只读工具,确认能返回结果,再接入正式流程。

典型使用场景

3 类
按分类与标签整理的常见用法,具体以官方文档为准
云资源查询

查实例、存储、账单等云上资源状态。

部署与发布

把部署、重启、回滚这类操作接进对话。

监控与告警

拉取监控指标,异常时让模型帮你判断。

选择建议

选之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 已适配 Claude Desktop、Claude,换客户端不用重新研究配置。
  • 资料标注无需鉴权,装好就能连。

先想清楚的地方

  • 没有现成安装命令,需要按官方文档自己填启动命令和参数。
  • 资料没标注工具数量,实际能力要装好后确认。
  • 版本更新与行为细节以官方仓库为准。

更适合

  • 需要在运维云端场景里让模型直接调用外部能力的人。
  • 已经在用 Claude Desktop、Claude 的用户,接入路径最短。
  • 想先快速试一下、能接受在本地或自有环境跑服务的人。

不太适合

  • 需要厂商级 SLA 和技术支持的生产环境。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
MCP 服务器工具数传输方式部署方式安装命令最后核验
Brain MCP Dcd6b298当前 未标注 未标注 未标注 无 2026-05-17
lobehub 未标注 stdio · 本地进程 本地部署 有 2026-04-11
TrendRadar 未标注 stdio · 本地进程 本地部署 有 2026-04-13
Ui Tars Desktop 8a825770 未标注 stdio · 本地进程 本地部署 有 2026-05-17
ruflo 未标注 stdio · 本地进程 本地部署 有 2026-04-13
FastGPT 未标注 stdio · 本地进程 本地部署 有 2026-04-11

服务器档案

运维云端
字段来自项目仓库与公开索引,中文说明由本站标注
提供方yonidavidson上游标注的维护方 yonidavidson
分类运维云端本站按上游分类展示
连接方式未标注上游未给出传输方式
鉴权方式无需鉴权远程服务建议用最小权限凭据
维护方社区 / 第三方核验等级 仅资料核对
包名 / 运行时未标注 · 未标注本地安装时用到的包标识与运行时
部署类型未标注上游给出的部署形态
最近核验2026-05-17上游最后核验时间

详细介绍

11.1 KB
来自项目仓库与上游目录的原文,未做删改

Brain MCP - Memory Store MCP Server

A Model Context Protocol (MCP) server that provides intelligent conversation memory storage using DuckDB. This server stores and retrieves the context of your last conversations with support for multiple storage backends including local files, S3, and Google Cloud Storage.

Features

  • 🧠 Smart Memory Storage: Automatically maintains context from the last 2 conversations
  • 💾 DuckDB Backend: Efficient data lake storage using DuckDB
  • ☁️ Multi-Backend Support: Store data locally or in the cloud (S3, GCS)
  • 🔄 Conversation Management: Track and retrieve multiple conversations
  • 🔍 Powerful Search: Search through memories by text, topics, or date range
  • 🌙 Automatic Memory Consolidation: Nightly worker using LLMs to consolidate short-term memories into long-term insights
  • 🤖 OpenAI-Compatible: Works with OpenAI API and compatible services (LocalAI, Ollama, etc.)
  • 🚀 Easy Integration: Standard MCP protocol for seamless integration

Installation```bash

npm install
npm run build


## Configuration

The server supports different storage backends through the `STORAGE_URL` environment variable:

### File-based Storage (Default)```bash
export STORAGE_URL="file://brain-memory.duckdb"

S3 Storage```bash

export STORAGE_URL="s3://my-bucket/brain-memory"
export AWS_ACCESS_KEY_ID="your-access-key"
export AWS_SECRET_ACCESS_KEY="your-secret-key"
export AWS_REGION="us-east-1"


### Google Cloud Storage```bash
export STORAGE_URL="gcs://my-bucket/brain-memory"
export GOOGLE_APPLICATION_CREDENTIALS="/path/to/credentials.json"

Memory Consolidation (Optional but Recommended)

The server includes an automatic nightly worker that consolidates short-term memories into long-term insights using an LLM:```bash

Required for consolidation

export OPENAI_API_KEY="your-api-key"

Optional: Use OpenAI-compatible services

export OPENAI_BASE_URL="https://api.openai.com/v1" # Or your compatible API
export OPENAI_MODEL="gpt-5-mini" # Default model

Optional: Customize consolidation schedule (cron format)

export CONSOLIDATION_SCHEDULE="0 0 * * *" # Default: midnight daily

Optional: Disable consolidation

export ENABLE_CONSOLIDATION="false" # Default: true


You can also copy `.env.example` to `.env` and configure your settings there.

## Usage

### Running the Server```bash
npm start

Or for development with auto-rebuild:```bash
npm run dev


### Configuring with Claude Desktop

Add this to your Claude Desktop configuration file:

**MacOS**: `~/Library/Application Support/Claude/claude_desktop_config.json`
**Windows**: `%APPDATA%/Claude/claude_desktop_config.json````json
{
  "mcpServers": {
    "brain-mcp": {
      "command": "node",
      "args": ["/path/to/brain-mcp/dist/index.js"],
      "env": {
        "STORAGE_URL": "file://brain-memory.duckdb",
        "OPENAI_API_KEY": "your-api-key-here",
        "OPENAI_MODEL": "gpt-5-mini"
      }
    }
  }
}

Available Tools

1. store_memory

Store a message in the conversation memory.

Parameters:
- role (required): The role of the message sender - "user", "assistant", or "system"
- content (required): The content of the message to store
- new_conversation (optional): Whether to start a new conversation (default: false)

Example:```json
{
"role": "user",
"content": "What is the weather like today?",
"new_conversation": false
}


### 2. `get_memory`

Retrieve stored conversation memory from the last N conversations.

**Parameters:**
- `conversation_count` (optional): Number of recent conversations to retrieve (default: 2, max: 10)

**Example:**```json
{
  "conversation_count": 2
}

Response:```json
{
"success": true,
"conversationCount": 2,
"currentConversationId": "conv-1234567890",
"conversations": [
{
"conversationId": "conv-1234567890",
"messageCount": 4,
"messages": [
{
"role": "user",
"content": "Hello!",
"timestamp": "2024-01-01T12:00:00.000Z"
}
]
}
]
}


### 3. `new_conversation`

Start a new conversation with a fresh conversation ID.

**Parameters:** None

**Example Response:**```json
{
  "success": true,
  "message": "New conversation started",
  "conversationId": "conv-1234567890"
}

4. get_long_term_memory

Retrieve consolidated long-term memories created by the consolidation worker.

Parameters:
- limit (optional): Number of memories to retrieve (default: 10, max: 50)

Example Response:```json
{
"success": true,
"count": 5,
"memories": [
{
"summary": "Discussion about implementing memory systems with DuckDB...",
"topics": ["DuckDB", "Memory Systems", "MCP Servers"],
"keyInsights": [
"DuckDB provides efficient storage for conversation data",
"Memory consolidation helps maintain context over time"
],
"consolidatedFrom": "Conversations from 2024-01-15",
"timestamp": "2024-01-16T00:00:00.000Z"
}
]
}


### 5. `consolidate_memory`

Manually trigger memory consolidation (normally runs automatically at night).

**Parameters:** None

**Example Response:**```json
{
  "success": true,
  "message": "Memory consolidation completed successfully"
}

6. search_memory

Search through both short-term and long-term memories using flexible filters.

Parameters:
- query (optional): Text to search for in conversation content, summaries, and insights
- topics (optional): Array of topics to filter long-term memories
- start_date (optional): Start date in ISO format (e.g., "2024-01-01T00:00:00.000Z")
- end_date (optional): End date in ISO format (e.g., "2024-12-31T23:59:59.999Z")
- memory_type (optional): Type of memory to search - "short-term", "long-term", or "both" (default: "both")
- limit (optional): Maximum number of results per memory type (default: 20, max: 100)

Example Request:```json
{
"query": "DuckDB",
"topics": ["databases", "storage"],
"memory_type": "both",
"limit": 10
}


**Example Response:**```json
{
  "success": true,
  "filters": {
    "query": "DuckDB",
    "topics": ["databases", "storage"],
    "dateRange": "all time",
    "memoryType": "both"
  },
  "shortTermResults": {
    "count": 5,
    "conversations": [
      {
        "conversationId": "conv-1234567890",
        "role": "user",
        "content": "Tell me about DuckDB performance...",
        "timestamp": "2024-01-15T14:30:00.000Z"
      }
    ]
  },
  "longTermResults": {
    "count": 3,
    "memories": [
      {
        "summary": "Discussion about DuckDB as a storage solution...",
        "topics": ["DuckDB", "databases", "storage"],
        "keyInsights": [
          "DuckDB provides excellent analytical query performance",
          "Embedded database with no server required"
        ],
        "consolidatedFrom": "Conversations from 2024-01-14",
        "timestamp": "2024-01-15T00:00:00.000Z"
      }
    ]
  }
}

Architecture

How Memory Consolidation Works

The Brain MCP implements a two-tier memory system inspired by human memory:

  1. Short-Term Memory (Conversations Table)
    - Stores recent conversation messages
    - Fast access for context in ongoing conversations
    - Automatically cleared after consolidation

  2. Long-Term Memory (Long-Term Memory Table)
    - Stores consolidated insights and summaries
    - Created by analyzing short-term memories with an LLM
    - Persists important information indefinitely

  3. Nightly Consolidation Process
    - Scheduled worker runs at midnight (configurable)
    - Retrieves all conversations from the current day
    - Analyzes them using the configured LLM model
    - Extracts key insights, topics, and creates summaries
    - Stores the consolidation in long-term memory
    - Clears short-term memory to free up space

This approach ensures:
- Recent context is always available
- Important information is preserved long-term
- Storage remains efficient
- Knowledge is organized and searchable

Storage Backends

The server uses DuckDB as its storage engine with support for:

  1. Local File Storage: Data is stored in a local DuckDB file
  2. S3 Storage: Data is persisted to Amazon S3 using DuckDB's httpfs extension
  3. GCS Storage: Data is persisted to Google Cloud Storage using DuckDB's httpfs extension

Data Schema

Short-Term Memory (Conversations):```sql
CREATE TABLE conversations (
id VARCHAR PRIMARY KEY,
timestamp BIGINT NOT NULL,
role VARCHAR NOT NULL,
content TEXT NOT NULL,
conversation_id VARCHAR NOT NULL
);


**Long-Term Memory:**```sql
CREATE TABLE long_term_memory (
  id VARCHAR PRIMARY KEY,
  timestamp BIGINT NOT NULL,
  summary TEXT NOT NULL,
  topics TEXT NOT NULL,          -- JSON array
  key_insights TEXT NOT NULL,    -- JSON array
  consolidated_from TEXT NOT NULL
);

Memory Management

  • The server automatically tracks conversations using unique conversation IDs
  • By default, it maintains and can retrieve the last 2 conversations
  • You can retrieve up to 10 recent conversations using the get_memory tool
  • Messages are ordered by timestamp for accurate conversation reconstruction

Development

Project Structure```

brain-mcp/
├── src/
│ ├── index.ts # Main MCP server implementation
│ ├── storage.ts # Storage manager with DuckDB integration
│ └── consolidation.ts # Memory consolidation worker
├── dist/ # Compiled JavaScript (generated)
├── package.json
├── tsconfig.json
├── .env.example
└── README.md


### Building```bash
npm run build

Type Checking

The project uses TypeScript with strict type checking enabled.

Environment Variables

Variable Description Default
Storage Configuration
STORAGE_URL Storage backend URL file://brain-memory.duckdb
AWS_ACCESS_KEY_ID AWS access key (for S3) -
AWS_SECRET_ACCESS_KEY AWS secret key (for S3) -
AWS_REGION AWS region (for S3) us-east-1
GOOGLE_APPLICATION_CREDENTIALS GCS credentials path -
Consolidation Configuration
OPENAI_API_KEY OpenAI API key (required for consolidation) -
OPENAI_BASE_URL OpenAI-compatible API endpoint https://api.openai.com/v1
OPENAI_MODEL Model to use for consolidation gpt-5-mini
CONSOLIDATION_SCHEDULE Cron schedule for consolidation 0 0 * * * (midnight)
ENABLE_CONSOLIDATION Enable/disable auto-consolidation true

Use Cases

Basic Usage

Store and retrieve recent conversations for immediate context.

Long-Term Knowledge Base

Automatically build a searchable knowledge base of important insights and information from all your conversations.

Multi-User Systems

Use different storage backends for different users or contexts.

Offline-First with Cloud Sync

Use local storage for fast access, with periodic sync to cloud storage for backup and sharing.

Custom Consolidation Schedules

Run consolidation multiple times per day, weekly, or on-demand based on your needs.

License

MIT

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

常见问题

6 问
答案由本站依据条目字段整理,不引用上游原文
Brain MCP Dcd6b298 怎么安装?

资料没有给出安装命令,请到官方仓库或文档按说明接入;这类条目多数是生态应用,或尚未发布可安装包。

Brain MCP Dcd6b298 怎么在 Claude Desktop、Claude 里配置?

它的传输方式未标注,在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,配置才会生效。

Brain MCP Dcd6b298 收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于服务本身和背后的接口额度,正式使用前按官方说明确认。

Brain MCP Dcd6b298 需要 API Key 吗?

资料标注为无需认证,装好就能连;如果实际调用时报鉴权错误,按官方文档补上凭据。

Brain MCP Dcd6b298 支持哪些客户端?

资料标注支持 Claude Desktop、Claude;其他客户端只要能配置 MCP Server,也能用同样的命令或地址接入。

Brain MCP Dcd6b298 能提供哪些能力?

资料显示能力标签包括 对话管理、TypeScript、Claude。具体工具清单以官方文档为准。

热门专题

按标签挑服务器
全部专题 →

服务器信息以作者仓库与官方文档为准;分类、传输方式与鉴权方式的中文说明为中文整理,最近核验 2026-05-17。