- fresh computer or VM with:
- macos 14.0+ (m series) or windows 11+ (x86_64)
- minimum 8gb ram
- minimum 30gb free disk space
- stable internet connection
- openai api key or anthropic api key or ollama running on your machine
- access to youtube
- screen device
- input audio device
- output audio device
- record your screen before starting the test
- download latest screenpipe from https://web.crabnebula.cloud/mediar/screenpipe/releases
- install and launch the app
- skip the onboarding flow
-
check recording status
- go to status page
- ensure dev mode is off
- click stop then start
- status should turn green within 60 seconds
-
configure ai
- navigate to settings > ai settings
- enter openai api key (or anthropic api key or ollama running on your machine)
- save settings
-
test search functionality
- go to main page
- enable search
- type "screenpipe" in search bar
- set time range to last 30 minutes
- press enter
- request summary
- verify summary accurately reflects app usage
-
play test video
- open https://youtu.be/UF8uR6Z6KLc?t=117
- play for exactly 120 seconds
-
verify transcription
- return to screenpipe main page
- click on meeting
- open transcript
- request summary
- verify accuracy of transcription and summary
- verify both your input and output audio devices have been used (2 check boxes should be checked with the name of your devices)
-
close application
- right click screenpipe icon in system tray
- select quit
-
verify process cleanup
- open activity monitor (macos) or task manager (windows)
- search for "screenpipe"
- verify no screenpipe processes running
- search for "bun"
- verify no bun processes running
- installation successful
- recording status works
- ai configuration successful
- search and summary functional
- meeting transcription accurate
- audio devices used
- clean process exit
- share screen recording
report any failures or unexpected behavior to https://github.com/mediar-ai/screenpipe/issues