https://github.com/activitywatch/activityscraper
Scrapes the activity the user is experiencing
https://github.com/activitywatch/activityscraper
Last synced: 11 months ago
JSON representation
Scrapes the activity the user is experiencing
- Host: GitHub
- URL: https://github.com/activitywatch/activityscraper
- Owner: ActivityWatch
- Created: 2017-08-26T00:35:58.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2017-08-26T02:28:59.000Z (over 8 years ago)
- Last Synced: 2024-10-30T00:54:46.457Z (about 1 year ago)
- Size: 1000 Bytes
- Stars: 6
- Watchers: 4
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
ActivityScraper
===============
*Scrape everything, leave nothing behind.*
## About
Alternative names: *Replaymember*, *Selfpast*
If ActivityScraper is given audio and video recordings, it can annotate them and provide playback of them. We can combine this with ActivityWatch and that **enables annotated recordings of all computer activity**. Using this we can build a complete story of what happened.
But perhaps more amazingly: we could train ML/AI with the data (since it's labeled) to auto-annotate (using image classifiers and deep learning).
This is an application enabled by ActivityWatch.
## Issues
- Storage
- Lowering resolution and quality (the most basic solution)
- Can be mediated using a high compression rate (but not an option since it needs to be real-time and energy efficient)
- Work
- When is the right time? Probably not a good time for quite a while, but cool stuff.
## PoC
- Doesn't do recording itself
- Just annotates given screen and/or audio recordings.
- Makes those recordings browsable/viewable in some rudimentary fashion
## Future
- Nice playback of the recordings
- Might be able to do recording itself if nice libraries are available
- Try to build auto-annotators using AI/ML (sounds like a possible task for deep learning)
## Uses
- Archive and label playing using annotations (would work against such horrors as WebDRM)
- You would have a searchable copy of every bit of video and audio you played on your computer.
## Far Future
- Maybe it could even understand VR content one day.