Analyze the user's computer usage patterns using raw SQL queries on their local Screenpipe database. Use when the user asks about their most used apps, websites, screen time, productivity, meeting time, typing habits, context switching, or any aggregate statistics about their activity.
Draft disabled Screenpipe automation bundles from an approved opportunity. Use when turning a passive suggestion into a reviewable `pipe.md` plus `automation.json` bundle.
Detect repeated workflows from Screenpipe activity and turn them into evidence-backed automation opportunities. Use for passive end-of-day suggestions, repeated work detection, or cross-device automation discovery.
Query structured UI elements (accessibility tree nodes and OCR text blocks) captured from the user's screen. Use when the user asks about specific buttons, text fields, UI components, links, or needs precise structural information about what was on screen.
Export videos, extract audio, and edit media clips from the user's Screenpipe recordings. Use when the user asks to export, download, clip, or create videos/audio from their screen recordings.
Create, install, and manage screenpipe pipes — scheduled AI agents that run automatically on a cron schedule. Use when the user asks to create a pipe, schedule a prompt, automate a task, set up recurring summaries, or turn a chat into a pipe.
Re-transcribe audio recordings with a different engine, custom vocabulary, or prompt. Use when the user wants to improve transcription quality, fix bad transcriptions, or re-process audio from a specific time range.
Search the user's screen recordings, audio transcriptions, and UI interactions via the local Screenpipe API at localhost:3030. Use when the user asks about their screen activity, meetings, apps they used, what they saw/heard, or anything about their computer usage history.