Screenpipe tutorial: recording, AI setup and scheduled pipes
Learn Screenpipe's recording checks, search and AI settings, then create a scheduled pipe that summarizes your day at 6 p.m.
Player not loading? Watch on YouTube
This Screenpipe walkthrough covers recording checks, searching captured activity and using that history with AI. The demonstration starts on Mac, where the speaker checks the recording indicator, audio devices and permissions. For builds of the open source version, he notes that a permission entry may need to be removed and added again.
The timeline lets users browse past activity, search recordings and inspect conversation transcripts alongside apps and visited websites. The speaker demonstrates connections to Claude and Codex, then explains AI presets. These include Screenpipe Cloud and locally running models.
Pipes are scheduled AI tasks that use Screenpipe recordings. The example asks for a daily summary at 6 p.m., then tests the pipe and refines it through feedback. The speaker also shows app connections and custom MCP tools for supplying additional context or performing actions.
The meeting section covers background transcription, a focused notes view, AI summaries and speaker reassignment. The closing sections show the recording data folder and support options. The walkthrough covers both locally stored recordings and cloud AI connections, so the demonstrated setup includes more than offline processing.
Design Blog
