how it works
screenpipe runs a local API onlocalhost:3030. Apple Shortcuts can call this API to search your screen history, get recent activity, or trigger pipes — all through Siri voice commands or automations.
Siri Shortcuts integration
search your screen with Siri
create a shortcut that queries screenpipe:- open Shortcuts app on macOS
- create a new shortcut
- add Get Contents of URL action:
- URL:
http://localhost:3030/search?q=QUERY&limit=5&content_type=ocr - method: GET
- URL:
- add Get Dictionary Value to parse the JSON response
- add Show Result or Quick Look to display results
example: “what was I working on?”
create a shortcut named “what was I working on”:- Get Contents of URL:
http://localhost:3030/search?content_type=ocr&limit=10 - Get Dictionary Value: key
data - Repeat with Each → extract
content.textandcontent.app_name - Show Result: display the text
example: get meeting transcriptions
- Get Contents of URL:
http://localhost:3030/search?content_type=audio&limit=20 - parse and display transcriptions
Apple Shortcuts automations
combine screenpipe with Shortcuts automations:| automation | how |
|---|---|
| daily summary at 6pm | time-based trigger → query screenpipe → format summary → send notification |
| focus mode logging | when Focus turns on → log current activity from screenpipe |
| meeting notes on calendar end | calendar event ends → query audio transcriptions → save to Notes |
tips
- screenpipe must be running for the API to respond
- all data stays local — Shortcuts queries
localhostonly - combine with Ask ChatGPT action in Shortcuts for AI-powered summaries
- use
start_timeandend_timeparams to scope queries to specific time ranges
vs Apple’s built-in screen time
| feature | screenpipe | Apple Screen Time |
|---|---|---|
| full text search | ✅ OCR on every frame | ❌ app usage only |
| audio transcription | ✅ local Whisper | ❌ |
| AI integration | ✅ any LLM | limited to Apple Intelligence |
| API access | ✅ full REST API | ❌ |
| cross-platform | ✅ macOS, Windows, Linux | ❌ Apple only |
| data export | ✅ SQLite, API | ❌ |
next steps
- install screenpipe
- MCP server — connect to Claude and Cursor
- pipes — scheduled AI agents on your data