imprint-mcp-server

imprint-mcp-server

Enables querying local meeting notes, calendar events, email metadata, and daily digests from Imprint's database via natural language through Claude Desktop and other MCP clients.

Category
访问服务器

README

imprint-mcp-server

A Model Context Protocol server that exposes Imprint's local data — meeting notes, calendar events, email metadata, daily digests — to Claude Desktop, Claude Code, and any other MCP-compatible client.

Read-only by design. The server never writes to Imprint's database; Imprint remains the source of truth for sync, classification, and writes.

What you can ask Claude

"What did we agree on in the Loyco meeting last week?"

"List my meetings tomorrow with online links."

"How many emails did I get from acme.no in April?"

"Summarize what I worked on this Tuesday."

Claude calls the appropriate tool(s), gets structured JSON back, and answers with full context from your own local Imprint data.

Tools

Name Purpose
search_meeting_notes Keyword search over titles, summaries, AI summaries, project context.
get_meeting_note Full content of one note + optional full transcript.
list_calendar_events All events on a given date, with online-meeting URLs.
search_emails Filter email metadata by query / sender domain / direction / date range. (Imprint never stores email bodies.)
get_daily_digest Imprint's pre-generated work-hours/focus/apps summary for a date.

Install

npm install -g imprint-mcp-server

Or use it ad-hoc via npx (recommended — no global install needed):

// ~/Library/Application Support/Claude/claude_desktop_config.json
{
  "mcpServers": {
    "imprint": {
      "command": "npx",
      "args": ["-y", "imprint-mcp-server"]
    }
  }
}

Restart Claude Desktop. You'll see Imprint show up in the tools menu, and Claude will call its tools whenever a question matches.

Custom database location

By default the server looks at the standard Imprint path:

~/Library/Application Support/Imprint/imprint.db

Override via env var if Imprint's DB lives elsewhere:

{
  "mcpServers": {
    "imprint": {
      "command": "npx",
      "args": ["-y", "imprint-mcp-server"],
      "env": { "IMPRINT_DB_PATH": "/path/to/imprint.db" }
    }
  }
}

Privacy

  • The server runs locally on your machine. The MCP transport is stdio — there's no network listener.
  • The SQLite connection is opened read-only.
  • Tool results travel from the MCP server to Claude Desktop via stdio, then Claude Desktop sends them to Anthropic's API for the model to read. So: whatever you tell Claude to look up is sent to the model, just like any other Claude conversation. Use the same judgment you would if you pasted the data into a chat.
  • Imprint stores only email metadata (sender domain, subject, timestamp) — no message bodies. The MCP server can't surface what isn't there.

Development

git clone <repo>
cd imprint-mcp-server
npm install
npm run build
npm start

Smoke-test by piping JSON-RPC at it:

(
  echo '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"smoke","version":"1"}}}'
  echo '{"jsonrpc":"2.0","method":"notifications/initialized"}'
  echo '{"jsonrpc":"2.0","id":2,"method":"tools/list"}'
  sleep 0.3
) | node dist/index.js

Releasing

Releases are published to npm via GitHub Actions using OIDC trusted publishing — no long-lived tokens are stored in the repo or secrets.

To cut a release:

# 1. Bump version in package.json
npm version patch  # or minor / major

# 2. Push the tag
git push --follow-tags

The Release to npm workflow runs on tag push, builds, and publishes with --provenance (a signed SLSA attestation linking the tarball to the exact commit + run — visible on the npm package page).

One-time npm setup (already done for imprint-mcp-server)

On npmjs.com → Packages → imprint-mcp-server → Settings → Trusted Publishers, the following GitHub Actions config is registered:

Field Value
Organization PTrobe
Repository imprint-mcp-server
Workflow filename release.yml
Environment name (empty)

This means npm will accept publishes from .github/workflows/release.yml on the PTrobe/imprint-mcp-server repo without any secret. Tokens with 90-day expirations stop being a maintenance burden.

Roadmap

  • [ ] Tasks tool (requires Imprint to cache tasks in DB; today they're fetched live per provider)
  • [ ] Activity-events search (window/app tracking)
  • [ ] Resources (vs. tools) for browsing notes by folder
  • [ ] Write tools — e.g. create_meeting_note, complete_task

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选