different-ai/openwork-hub · Archived

screenpipe

Screenpipe records your screen and microphone 24/7. Search your screen recordings and audio transcriptions via REST API and CLI.

First seen Mar 13, 2026

Installation

$ npx skills add different-ai/openwork-hub --skill screenpipe

Stronger alternatives

This repository is archived — consider an actively maintained alternative.

Also in this package

Other skills from different-ai/openwork-hub · top by installs.

npx skills add different-ai/openwork-hub

Browse all from different-ai/openwork-hub

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 26
Default branch main
Open issues 0
Status Archived

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 3,832 B
  • docs SUMMARY.md 146 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 3 installs

SKILL.md

Quick Usage

Search via API

# Search for text in screen recordings
curl "http://localhost:3030/search?q=meeting&limit=10"

# Search audio transcriptions only
curl "http://localhost:3030/search?q=budget&content_type=audio"

# Search with time range (ISO 8601)
curl "http://localhost:3030/search?q=invoice&start_time=2024-01-28T09:00:00Z&end_time=2024-01-28T17:00:00Z"

# Filter by app
curl "http://localhost:3030/search?q=code&app_name=Visual%20Studio%20Code"

# Get recent activity (no query = all content)
curl "http://localhost:3030/search?limit=50"

Search via CLI

# Basic search
screenpipe search "meeting notes"

# Search with filters
screenpipe search "api key" --app Chrome --limit 20

# Search audio only
screenpipe search "what did they say" --type audio

# Search last hour
screenpipe search "error" --from "1 hour ago"

API Response Format

{
  "data": [
    {
      "type": "OCR",
      "content": {
        "text": "Meeting at 3pm with Sarah",
        "timestamp": "2024-01-28T14:32:00Z",
        "app_name": "Google Chrome",
        "window_name": "Calendar - Google Chrome"
      }
    },
    {
      "type": "Audio",
      "content": {
        "transcription": "Let's discuss the Q1 budget",
        "timestamp": "2024-01-28T15:00:00Z",
        "device_name": "MacBook Pro Microphone"
      }
    }
  ],
  "pagination": {
    "limit": 10,
    "offset": 0,
    "total": 142
  }
}

API Reference

GET /search

Parameter Type Description
q string Search query (optional - omit for all content)
content_type string ocr, audio, ui, or all (default: all)
limit int Max results (default: 10, max: 100)
offset int Pagination offset
start_time string ISO 8601 UTC timestamp
end_time string ISO 8601 UTC timestamp
app_name string Filter by application
window_name string Filter by window title
min_length int Minimum content length
max_length int Maximum content length

GET /health

Check if Screenpipe is running:

curl http://localhost:3030/health

Other Endpoints

Endpoint Description
GET /audio/list List audio devices
GET /tags List all tags
POST /tags/:id Add tag to content
DELETE /tags/:id/:tag Remove tag

CLI Reference

# Start screenpipe daemon
screenpipe

# Start with specific settings
screenpipe --fps 1 --audio-chunk-duration 30

# Disable audio
screenpipe --disable-audio

# Use specific audio device
screenpipe --audio-device "MacBook Pro Microphone"

Common Gotchas

  • Screenpipe must be running locally on port 3030.
  • Audio transcription requires microphone permissions.
  • Check logs in $HOME/.screenpipe
  • Screen capture requires screen recording permissions (macOS: System Preferences → Privacy & Security → Screen Recording).
  • URL capture and shortcuts requires accessibility permissions
  • Search results are paginated - use offset for more results.

First-Time Setup

1. Install Screenpipe

Download the app: https://screenpi.pe/onboarding

Or build from source:

git clone https://github.com/mediar-ai/screenpipe
cd screenpipe
cargo build --release

2. Start Screenpipe

screenpipe

Or launch the desktop app.

3. Verify it's running

curl http://localhost:3030/health

Resources