Back to Catalog

screenpipe

mediar-ai/screenpipe
🔗 Latest commit:c2a5a00
🕒 Updated:Aug 3, 2025, 09:01 PM
TypeScript
AI Tools

AI app store powered by 24/7 desktop history. open source | 100% local | dev friendly | 24/7 screen, mic recording

MCP Trust Score
Based on our comprehensive evaluation criteria
🤖 Evaluated by gemini-2.5-flashFix
Trust Score66/100
GitHub Metrics
Repository statistics and activity
⭐ GitHub Stars:15381
👥 Contributors:78
📋 Total Issues:205
📦 Has Releases:Yes
🔧 Has CI/CD Pipeline:Yes
Configuration
Configuration example extracted from README.md for Claude Desktop and other clients.
🤖 Evaluated by gemini-2.5-flashFix
{
  "mcpServers": {
    "screenpipe": {
      "command": "screenpipe",
      "args": [],
      "env": {}
    }
  }
}
MCP Protocol Support
Implemented MCP protocol features
🤖 Evaluated by gemini-2.5-flashFix
Tools:
Prompts:
Resources:
Sampling:
Roots:
Logging:
STDIO Transport:
HTTP Transport:
OAuth2 Auth:
Dependencies
160 dependencies
Libraries and frameworks used by this MCP server
🤖 Evaluated by gemini-2.5-flashFix
Add Quality Badge
Show your MCP trust score in your README
Trust Score Badge
[![Trust Score](https://archestra.ai/mcp-catalog/api/badge/quality/mediar-ai/screenpipe)](https://archestra.ai/mcp-catalog/mediar-ai__screenpipe)
README.md

English | 简体中文 | 日本語

logo

mediar-ai%2Fscreenpipe | Trendshift

[ screenpipe ]

AI app store powered by 24/7 desktop history

open source | 100% local | dev friendly | 24/7 screen, mic recording

[ recording reality, one pixel at a time ]

Download the Desktop App

YouTube Channel Subscribers

Join us on Discord X account Rewarded Bounties Open Bounties

Screenshot 2025-02-15 at 7 51 18 PM

image

https://github.com/user-attachments/assets/628c6c01-a580-4b21-bce9-3e7b186914a4

https://github.com/user-attachments/assets/973ee8e5-5240-4d36-83fe-d38c53efe6a9


news 🔥


how it works?

  • we record everything 24/7, 100% locally, uses 10% CPU, 4 GB ram, 15 gb/m
  • we index it into an api
  • dev build ai apps w user's full context, desktop native, nextjs, publish, monetize

why?

  • ai models are commoditized
  • ai is as good as its context
  • the most valuable context is all contained in your screen

get started

macos, linux:

curl -fsSL get.screenpi.pe/cli | sh

or on windows

iwr get.screenpi.pe/cli.ps1 | iex

then

screenpipe

make sure to allow permissions on macos (screen, mic)

create plugins

bunx --bun @screenpipe/dev@latest pipe create

screenpipe has a plugin system called "pipe" which lets you create desktop app in nextjs in a sandboxed environment within our Rust code, read more

you can then publish these to our store and make money:

cd foo
bunx --bun @screenpipe/dev@latest pipe register --name foo [--paid --price 50] # subscription
bun run build
bunx --bun @screenpipe/dev@latest pipe publish --name foo

community

star history

Star History Nov 24 2024

contributing

contributions are welcome! if you'd like to contribute, please read CONTRIBUTING.md.

Rewarded Bounties Open Bounties
screenpipe MCP Server | Documentation & Integration | Archestra