AI Second Brain
screenpipe is a local-first, private AI agent memory that captures what you've seen, heard, and done, giving AI agents the context and information they never had, so they never start from scratch.
— local first, on your machine —
used by engineers and researchers at
Google
NVIDIAIntelShopifyAtlassian
Datadog
Adobe
Salesforce
Thomson Reuters
Broad Institute
MITStanfordHarvard
Planet Labs
WorktraceGoogleNVIDIAIntelShopifyAtlassianDatadogAdobeSalesforceThomson ReutersBroad InstituteMITStanfordHarvardPlanet LabsWorktraceGoogleNVIDIAIntelShopifyAtlassianDatadogAdobeSalesforceThomson ReutersBroad InstituteMITStanfordHarvardPlanet LabsWorktraceGoogleNVIDIAIntelShopifyAtlassianDatadogAdobeSalesforceThomson ReutersBroad InstituteMITStanfordHarvardPlanet LabsWorktraceOne focused look at workflow memory, company knowledge, and the automations they unlock.
Search across screen, audio, apps, and handoffs without rebuilding the context.
Pick an agent. See what triggered it, what it read, what it did, and where the result went.
When a meeting ends, write a summary + action items and push to your notes. Runs on-device.
---
name: meeting-notes
trigger: meeting_ended
model: local
---
When a meeting ends, write:
- 3-bullet summary
- Action items per person
- Decisions + next steps
Save to ~/notes/{date}_{app}.md
Click Simulate meeting end to see the notification.
Connect once. Pipes use them automatically.
Connect once. Your pipes use them automatically — Gmail for meeting follow-ups, HubSpot for CRM updates, Slack for notifications.
Exclude any app, window, or URL. Those are dropped at the source. Everything else is scrubbed by a screenpipe model on your machine that removes cards, SSNs, and keys before anything is saved.
Apps, windows, and URLs you exclude never reach the disk. Everything else is scrubbed on your machine — flip a switch to see what changes.
Everything sensitive is scrubbed right on your machine — 0 bytes leave your device.
An AI second brain is a memory layer that stores your personal context, so your AI can retrieve what you've already seen, read, and done instead of starting from scratch. screenpipe is exactly that: a local-first AI agent memory, and the AI second brain that builds itself.
The usual way to build an AI second brain is manual. You write notes into Obsidian, wire up Claude Code to read them, and hope you captured the parts that matter. It works, but it only remembers what you remembered to write down, and it becomes one more thing to maintain.
screenpipe skips the busywork. It records your screen and audio in the background, extracts text with OCR and on-device transcription, and turns your everyday work into a searchable second brain that Claude Code, Cursor, and ChatGPT can query. No vault to maintain, and no re-explaining your work.
Ready to give your AI context?