{"id":105615,"title":"Screenpipe: Record how you work and turn that into agents","tagline":"Record how you work and turn that into agents","body":"screenpipe records what you do, see and hear on your computer, makes it searchable, and gives agents the context to understand how you actually work. The project has 20K+ GitHub stars.\n\n\u003chttps://youtu.be/c1jV6E9pyug\u003e\n\nMost AI agents start every task with partial context.\n\nThey can read the prompt and apps you connect. They do not remember the meeting where a decision was made, the tab you opened last Tuesday, the exact steps you took to fix a problem, or the edge case you discovered while doing the work.\n\nscreenpipe runs quietly in the background and captures screen and audio context on your machine. Then you can ask:\n\n* maintain a second brain for every person, projects, tasks, meetings\n\n\n* turn the workday into a daily recap and time breakdown for standups or client billing\n* maintain a personal CRM from the people, conversations, and follow-ups seen during the day\n* find commitments mentioned in meetings or messages that never made it into a task list\n* analyze repeated work and produce a list of workflows worth automating\n* save prompts and context from Claude, Codex, and ChatGPT into a searchable journal\n\nAgents can use screenpipe as long-term memory, so they work from what actually happened instead of only what you remembered to paste into a prompt.\n\nThe same foundation also works across team machines. Teams can map repeated workflows, generate SOPs, find bottlenecks, and identify automation opportunities across the tools where work actually happens.\n\nscreenpipe is open source and local-first. You control what gets recorded, can exclude apps, windows, and URLs, and choose where your data lives.\n\nI’ve been building my own AI second brain since 2020, but it could only remember what I explicitly saved. I built screenpipe so my AI could understand what I was actually doing across my screen and conversations. Other people began using it for the same reason. Today, we’re turning that foundation into context for personal and team agents.\n\n**Our ask**\n\nTry screenpipe and comment with the agent you would build if it could remember everything you’ve worked on.\n\nTry screenpipe: \u003chttps://screenpipe.com\u003e","slug":"RTT-screenpipe-record-how-you-work-and-turn-that-into-agents","created_at":"2026-07-14T14:38:00.824Z","updated_at":"2026-07-22T13:24:34.127Z","total_vote_count":20,"url":"https://www.ycombinator.com/launches/RTT-screenpipe-record-how-you-work-and-turn-that-into-agents","share_image_url":"//bookface-static.ycombinator.com/assets/ycdc/yc-og-image-c440a0ad1dacfb86eeeb343717479cc54d256614449b4ef719977a0a451f8bc8.png","company":{"id":32259,"name":"screenpipe","slug":"screenpipe","url":"https://screenpipe.com","logo":"https://bookface-images.s3.amazonaws.com/small_logos/144e7207ec725036f43cc9a6b577a8320151e103.png","batch":"Summer 2026","industry":"B2B","tags":["SaaS","B2B","Enterprise","Enterprise Software","AI"],"search_path":"https://bookface.ycombinator.com/company/32259"}}