Littlebird Builds a Work Memory From Screen and Meetings

My notes

Summary

Littlebird is a macOS and Windows AI assistant that passively watches the active screen and listens to meetings to build a running memory of your work, then answers questions, drafts documents, and schedules routines without anyone having to paste context. Early users position it as the spiritual successor to Rewind (which its founder wound down), built by a team with Sentieo roots.

Key Insight

• Core mechanism: it doesn’t rely on manual context-pasting. It watches the active window (documents, browser tabs) and listens during meetings, building a continuous work memory automatically, with no setup step required.
• Deep integrations (hundreds) span project management (Notion, Linear, ClickUp, Jira and Confluence), CRM and sales (Intercom, Outreach, Close), finance (Mercury, Ramp, Stripe, PayPal), meetings (Fireflies, Granola, Krisp, Circleback), and marketing and analytics (Ahrefs, Mixpanel, PostHog). The pitch is that combining passive screen capture with deep tool integration removes the “re-explain your own work” tax that generic chat assistants impose.
• Explicit privacy and trust framing to counter the obvious objection to screen-watching software: SOC 2 certified, GDPR and CCPA compliant, data encrypted at rest and in transit, no training on user data, no data sale, and granular deletion (all data, or just the last hour or day).
• Data is captured locally but stored and processed in the cloud on AWS, so this is not a fully local on-device memory model despite the “sees your screen” framing.
• Distribution model: free to start, a $15/month discounted plan for verified students (.edu), and a companion iOS and Android app for on-the-go queries when away from the desktop where the actual capture happens.
• Testimonials from named operators frame it as reducing “the overhead of remembering, retrieving, and re-explaining your own work”, a distinct value proposition from RAG-over-uploaded-docs assistants.