screenpipe
Local-first workflow memory for AI agents
About
screenpipe captures and indexes screen activity, audio, meetings, and workflows locally on macOS, Windows, and Linux. Users can search their work history directly or let Claude, Codex, OpenClaw, Hermes, and other MCP-compatible agents query selected context. The result is practical workflow memory: cited SOPs, task-mining insights, meeting recall, and automation candidates. Core capture and search can stay entirely on-device, with app exclusions, retention controls, and optional PII redaction.
Pricing
Basic $25/month; Business $50/seat/month; Enterprise $150/seat/month. Source available for free personal, non-commercial use.
Submitted by
Louis Beaumont
More in Agent Infrastructure

Ollagraph
Web Infrastructure for AI Agents

Qarinah
Cited project memory for coding agents with 98.71% less repeated context
Wuniq KE
Persistent, structured project memory for every MCP-compatible AI
Built something?
Submit your tool to get featured on our daily launches and a shoutout on social.
Submit your tool