I used a $2/month AI to build my personal encyclopedia — here's the setup

Last week, 'Personal Encyclopedias' hit #1 on Hacker News with 676 points. The comments were full of developers sharing their note systems, Obsidian vaults, Notion databases, and Roam graphs.

But almost nobody mentioned AI.

Here's how I use a $2/month AI assistant to build and query my personal knowledge base.

The problem with personal knowledge management

PKM tools are good at storing information. They're terrible at retrieving it in a useful way.

You write a note in 2022 about a React pattern you discovered. In 2024, you need that pattern again. Do you remember the note exists? Do you remember what you titled it? Can you find it in the search?

Most people can't. The note dies.

What AI changes about PKM

AI doesn't forget. AI doesn't care about file names or folder structures. AI understands meaning, not just keywords.

The workflow I use:

1. Write notes in plain text (whatever format you prefer)
2. Paste relevant context into the AI conversation
3. Ask questions in natural language
4. Get synthesized answers with connections you didn't notice

Example prompt:

"I'm working on a caching problem in Node.js. What did I learn about Redis last year? What patterns have I used before?"

The AI finds the thread, connects the dots, suggests the pattern.

The $2/month constraint

Here's the thing about expensive AI subscriptions: they change how you use the tool.

At $20/month, you feel pressure to use it constantly to justify the cost. You start using it for things it's bad at. You get disappointed. You cancel.

At $2/month, the pressure disappears. You use it when it helps. You build a habit naturally.

I built SimplyLouie specifically because I wanted Claude's intelligence (Anthropic's model, which I trust) at a price that doesn't create psychological pressure.

The actual setup

Step 1: Dump your context

Before asking a PKM question, paste the relevant notes into the conversation:

Here's what I have on Redis from my notes:
[paste notes]

And here's my current Node.js caching problem:
[describe problem]

What connections do you see?

Step 2: Ask for synthesis, not search

The AI isn't a search engine. It's a thinking partner. Ask it to synthesize:

"Looking at everything I've written about distributed systems, what patterns do I keep returning to? What am I missing?"

Step 3: Write the output back to your notes

The AI's synthesis becomes a new note. The cycle continues. Your knowledge compounds.

Real example from last month

I was debugging a performance issue in a side project. I asked the AI:

"I have these three performance problems: [described them]. What's the root pattern?"

The AI identified that all three were symptoms of the same N+1 query pattern — something I knew abstractly but hadn't connected to my specific code.

Two-minute conversation. Two-week debug session avoided.

The PKM tools I actually use with this setup

• Obsidian for long-form notes (plain markdown, local files)
• Logseq for daily notes and quick captures
• SimplyLouie ($2/month) for the AI synthesis layer

I've stopped thinking of AI as a separate tool. It's the query layer on top of everything else.

Why $2/month matters for this workflow

Personal knowledge management is a daily habit. It only works if you do it every day.

A $20/month subscription creates a psychological barrier. You start asking "is this worth a dollar of my subscription?" before every conversation.

At $2/month — less than $0.07/day — the question disappears.

The best PKM tool is the one you actually use.

---

SimplyLouie is a $2/month Claude-powered AI assistant. 7-day free trial, no commitment. 50% of all revenue goes to animal rescue. Try it at simplylouie.com.