An open source Second Brain system that captures, classifies, and connects your thoughts automatically using agentic AI.

What It Does

Send a message to a Slack bot. It figures out what to do with it. No apps to open, no categories to choose, no formatting decisions. Just capture and let the automation handle the rest.

Key Features:

  • Frictionless capture via Slack with zero cognitive overhead
  • AI-powered classification using AWS Bedrock AgentCore with behavioral learning
  • Intelligent linking that connects related thoughts across time
  • Automatic execution routing to external systems (OmniFocus, calendar, etc.)
  • Semantic search across your entire knowledge base
  • Git-backed storage with versioned Markdown files

Technical Architecture

Built on AWS serverless infrastructure:

  • Lambda + API Gateway for Slack webhook processing
  • AgentCore + Bedrock for intelligent message classification
  • CodeCommit for versioned knowledge repository
  • SQS + DynamoDB for reliable, idempotent processing
  • SES for routing actionable items to external systems

What I found using it

This is a personal project, and these are observations from my own use rather than measured benchmarks:

  • Zero-friction capture noticeably increased how often I recorded notes
  • Processing typically completed in a second or two from message to stored note, measured informally
  • Related ideas were linked without manual tagging
  • The architecture was composable and adapted to my existing workflow

The Vision

This represents the evolution from passive storage to active intelligence. These systems don’t just remember, they think with you. Future enhancements include multi-agent orchestration with specialized AI agents for research, synthesis, execution, and insights.

Built with AWS CDK, Lambda, AgentCore, and DynamoDB. A personal experiment in agentic AI and composable serverless architecture, not a production system.