Second Brain

The problem. I had too much information scattered across different sources to keep track of: articles, blog posts, news, and dense learning materials from MIT, Harvard, my PMP, CPMAI, and Change Management certifications. I needed a way to structure and organize this knowledge without getting locked into a proprietary ecosystem or spending hours on manual maintenance.

Why I built it. I wanted my knowledge to compound over time rather than decaying into digital clutter. By combining a local Obsidian markdown vault with developer tools like Claude Code and Antigravity, I created a system that easily reads my context, suggests improvements, and gets smarter the more I use it.

What it does. A suite of custom Python scripts maintains the vault by removing duplicates, generating a weekly digest, and auto-linking related concepts. It runs a News Curator Telegram bot that sends me summaries of topics I follow; replying to the bot automatically ingests the article. A voice pipeline transcribes mobile voice memos using a local LLM, preserving both my raw thoughts and a polished, cleaned-up version. It is also my central project command center, managing the feature backlogs, bugs, and research for all my active projects.

The benefits. Zero-overhead organization that actually gets better over time instead of defaulting to chaos. Deep linking mimics how our own brains connect ideas, allowing insights to cross-pollinate. The voice pipeline turns raw reflections into structured notes privately via local LLMs, serving as the source for many of the thoughts shared on this site.

Features

• Seamless integration with Claude Code and Antigravity for agentic workspace support
• Automated wiki linking, duplicate cleanup, and a weekly digest script
• Telegram-based News Curator with automated page creation and ingestion
• Voice pipeline utilizing a local LLM to transcribe and clean up mobile voice notes
• Central project management system tracking feature backlogs and research