kips
Personal data management MCP server that stores and queries notes, tasks, conversations, and auth in SQLite, enabling LLMs to interact with the database through natural language.
README
kips
Personal data management server. Integrates with Claude Desktop and other MCP clients.
Usage
kips drives a SQLite database and surfaces tools for LLMs as well as a basic CLI for interacting with the database.
You can store auth (passwords and usernames), notes, conversations, tasks (leftover work to do from last time) and of course tags.
Notes, conversations and tasks have join tables with tags under noteTag, conversationTag and taskTag.
Just run it normally first to init the database:
npx kips
Then ingest data:
## passwords in csv files
npx kips import --type auth ~/passwords.csv
## notes and tags
npx kips import --type note ~/note.txt --tag "big-deal another-tag"
## conversations in the past
npx kips import --type conversation ~/claudechat.txt --tag "therapy"
## agentic tasks that aren't done yet
npx kips import --type task ~/tasks.csv --tag "therapy revenge"
Integrate it into Claude Desktop to let Claude query, insert or update the database by running the config command:
npx kips config
You'll then be able to see the MCP server in Claude Desktop (usually after restart) with the hammer icon, and all applicable tables as resources you can "connect" with the plug icon.
Expected formats
The csv imports expect specific headers in order to be useful.
Auth
We expect a .csv file with url, username or email (both can exist, but email takes precedence), password and notes columns. All other columns are ignored.
Tasks
We expect a .csv file with objective, progressAssessment and completed columns. completed is a boolean.
Usage with Claude Desktop
All operations are specified for MCP clients and all schemas are exposed as attachable resources. If you have more complex files (screenshots, PDFs) you can get Claude to read and transcribe it direct to the database. At present, you'll need to attach all schemas you intend to use or Claude will have to look up the schema again before applying an operation.
Development
You'll want the Claude integration to use your local script.
## get it installed and built
npm run build
## now configure it to use the local build
node ./build/index.js config --debug
Likewise, when developing you can either npm link to use your local install as an npx package or just call the script direct:
## eg, using the prior example
node ./build/index.js import --type task ~/tasks.csv --tag "therapy revenge"
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。