Most important take away

Building a personal “OpenBrain” database with a visual front-end gives both you and your AI agents a shared, persistent memory that survives across sessions and tools. The key architectural insight is that the database table is the single source of truth — your agent reads and writes via MCP, and you read and write through a simple web interface, with no sync layer or middleware in between.

Chapter Summaries

The Problem: AI Memory Loss

Every new AI session starts from zero. OpenBrain solves this with a personal Supabase database connected to any AI via MCP server, but the original system was text-only through a chatbot “keyhole.”

Adding a Human Door

Layer a visual web interface over the same database tables so humans can scan, browse, and edit data visually while agents continue accessing it through MCP. No sync layer needed; both sides read/write the same rows.

Building & Deploying

Describe the desired interface to your AI (Claude, ChatGPT), have it generate a small web app, then deploy for free on Vercel. Bookmark it on your phone and it behaves like a native app.

Use Case: Household Knowledge

Log paint colors, shoe sizes, service dates, passwords, etc. into a table during normal AI conversations. Build a searchable, categorized visual interface over it.

Use Case: Professional Relationships

Track contacts, last interactions, and conversation context. Surface neglected relationships proactively. OpenClaw can autonomously flag connections going cold.

Use Case: Job Search Dashboard

Manage parallel workstreams (companies, contacts, applications, interviews, resume versions) across linked tables. Cross-reference conference notes with job postings to surface warm introductions.

Design Principles

Look for problems where (1) an agent can bridge time (maintenance schedules, deadlines), (2) an agent sees across categories (cross-referencing tables humans wouldn’t), and (3) humans decide while agents execute memory and pattern recognition.

Future-Proofing

Because everything is built on MCP, every new AI model automatically makes your system smarter. No vendor lock-in, no SaaS middlemen.

Summary

Actionable Insights

• Build visual dashboards over your OpenBrain tables by describing what you want to your preferred AI, having it generate a web app, and deploying it on Vercel for free.
• Start with household knowledge: log institutional knowledge (paint colors, passwords, service dates) during normal AI conversations and build a searchable interface.
• Track professional relationships in a table and let autonomous agents proactively flag neglected contacts.
• Create a job search dashboard linking contacts, companies, applications, and interviews to surface warm introductions and pipeline patterns.
• Identify agent-ready problems by looking for scattered information, events spread across time, and multiple categories needing cross-referencing.

Career Advice

For job seekers, the system can cross-reference conference notes with job postings to find warm introductions instead of cold applications. Track which resume versions perform best. Let autonomous agents catch expiring follow-up windows you would otherwise forget.