screenpipe

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Screenpipe — AI 屏幕记忆

Screenpipe — AI Screen Memory

24/7 记录屏幕内容和音频，让 AI 拥有「记忆」。用户可以回溯任何看过的内容、说过的话、用过的应用。所有数据本地存储，隐私优先。

Records screen content and audio 24/7 to give AI "memory". Users can trace back any content they have seen, words they have spoken, and apps they have used. All data is stored locally, with privacy as the first priority.

使用场景

Usage Scenarios

用户说「我昨天在哪个网页上看到那篇关于…的文章？」
用户说「上周开会时谁说了什么？」
用户说「今天我在电脑上花了多少时间在哪些应用上？」
用户说「帮我找一下之前看到的那个代码片段」
用户说「我这周总共写了多少代码？看了多少邮件？」
结合每日简报类 Skill（如已启用）自动生成基于真实屏幕活动的每日回顾

User asks: "Which webpage did I see that article about ... on yesterday?"
User asks: "Who said what during the meeting last week?"
User asks: "How much time did I spend on which apps on my computer today?"
User asks: "Help me find that code snippet I saw before"
User asks: "How much code have I written this week in total? How many emails have I read?"
Combine with daily briefing Skills (if enabled) to automatically generate daily reviews based on real screen activities

前置条件

Prerequisites

安装 Screenpipe：https://screenpi.pe/ （macOS / Windows / Linux）
启动 Screenpipe 后，本地 API 运行在
```
http://localhost:3030
```
Screenpipe 内置 MCP Server，通过 MCP 协议自动连接

Install Screenpipe: https://screenpi.pe/ (macOS / Windows / Linux)
After starting Screenpipe, the local API runs on
```
http://localhost:3030
```
Screenpipe has a built-in MCP Server, which automatically connects via the MCP protocol

执行方式

Execution Methods

通过 MCP 工具调用

Call via MCP Tool

Screenpipe MCP 提供以下核心能力：

Screenpipe MCP provides the following core capabilities:

搜索屏幕内容（OCR 文字）

Search screen content (OCR text)

工具: search_screen_content
参数:
  query: "搜索关键词"
  start_time: "2026-02-25T00:00:00Z"  # 可选，时间范围
  end_time: "2026-02-26T00:00:00Z"
  app_name: "Chrome"  # 可选，限定应用
  limit: 10

Tool: search_screen_content
Parameters:
  query: "search keyword"
  start_time: "2026-02-25T00:00:00Z"  # optional, time range
  end_time: "2026-02-26T00:00:00Z"
  app_name: "Chrome"  # optional, limit to specific app
  limit: 10

搜索音频转录

Search audio transcripts

工具: search_audio_transcripts
参数:
  query: "会议讨论内容"
  start_time: "2026-02-25T09:00:00Z"
  limit: 5

Tool: search_audio_transcripts
Parameters:
  query: "meeting discussion content"
  start_time: "2026-02-25T09:00:00Z"
  limit: 5

获取应用使用统计

Get app usage statistics

工具: get_app_usage
参数:
  start_time: "2026-02-25T00:00:00Z"
  end_time: "2026-02-26T00:00:00Z"

Tool: get_app_usage
Parameters:
  start_time: "2026-02-25T00:00:00Z"
  end_time: "2026-02-26T00:00:00Z"

直接 API 调用（备选）

Direct API Call (Alternative)

如果 MCP 不可用，可通过 HTTP API 访问：

bash

undefined

If MCP is unavailable, you can access via HTTP API:

bash

undefined

搜索屏幕 OCR 内容

Search screen OCR content

curl "http://localhost:3030/search?q=关键词&content_type=ocr&limit=10"

curl "http://localhost:3030/search?q=keyword&content_type=ocr&limit=10"

搜索音频转录

Search audio transcripts

curl "http://localhost:3030/search?q=关键词&content_type=audio&limit=5"

curl "http://localhost:3030/search?q=keyword&content_type=audio&limit=5"

获取最近活动

Get recent activities

curl "http://localhost:3030/search?limit=20&start_time=2026-02-25T00:00:00Z"

undefined

curl "http://localhost:3030/search?limit=20&start_time=2026-02-25T00:00:00Z"

undefined

典型工作流

Typical Workflows

回溯查找：

用户：我昨天下午看到一个很好的 Python 库，名字里有 pipe
→ 搜索 OCR 内容，时间限定为昨天下午
→ 返回匹配的屏幕截图和上下文
→ 告诉用户：你在 Chrome 中浏览了 GitHub 上的 xxx 项目（14:32）

会议回顾：

用户：今天上午的会议讨论了什么？
→ 搜索音频转录，时间限定为今天上午
→ 提取关键讨论点和行动项
→ 结构化输出会议摘要

时间追踪：

用户：我今天在各个应用上花了多少时间？
→ 获取应用使用统计
→ 生成时间分布报告（可结合图表生成类 Skill 生成图表）

Traceback search:

User: I saw a very good Python library yesterday afternoon, with pipe in its name
→ Search OCR content, limit time to yesterday afternoon
→ Return matching screenshots and context
→ Tell the user: You browsed the xxx project on GitHub in Chrome (14:32)

Meeting review:

User: What was discussed in this morning's meeting?
→ Search audio transcripts, limit time to this morning
→ Extract key discussion points and action items
→ Output structured meeting summary

Time tracking:

User: How much time did I spend on each app today?
→ Get app usage statistics
→ Generate time distribution report (can be combined with chart generation Skills to generate charts)

与其他 Skills 的协作

Collaboration with Other Skills

组合	效果
screenpipe + 每日简报类 Skill（如已启用）	基于真实屏幕活动生成每日回顾
screenpipe + meeting-insights-analyzer	自动回顾会议录音和屏幕共享内容
screenpipe + habit-tracker	基于真实应用使用数据追踪习惯
screenpipe + pomodoro	回顾专注时段内的实际工作内容

Combination	Effect
screenpipe + daily briefing Skills (if enabled)	Generate daily reviews based on real screen activities
screenpipe + meeting-insights-analyzer	Automatically review meeting recordings and screen sharing content
screenpipe + habit-tracker	Track habits based on real app usage data
screenpipe + pomodoro	Review actual work content during focus periods

输出规范

Output Specifications

搜索结果附上时间戳和来源应用
涉及屏幕内容时描述上下文（哪个应用、哪个页面）
音频转录标注说话人（如果 Screenpipe 提供了说话人标签）
时间统计使用表格/图表呈现
尊重隐私：不主动提及用户未询问的敏感内容

Attach timestamps and source apps to search results
Describe context when involving screen content (which app, which page)
Mark speakers for audio transcripts (if Screenpipe provides speaker tags)
Present time statistics in tables/charts
Respect privacy: Do not actively mention sensitive content that the user has not asked about