Ten Tabs and a Dream: How I Built a ‘Brain’ for Lenny’s Newsletter
· Mohammad Syed
Key takeaways
- Lenny Rachitsky released his archive (350+ posts and 300+ podcast transcripts) with a challenge to build something. I built Lenny’s Product Brain in a little over a week.
- It has Ask Lenny AI (RAG search with citations, powered by Gemini), a Knowledge Graph of 508 topics, 439 documents and 247 guests, Learning Paths, a PM Coach and a Content Library.
- Locking down the data to protect Lenny’s IP broke the whole app. The fix was controlled windows that give the UI just enough data and never the raw content.
- The gap between “content exists” and “knowledge is accessible” is bigger than most of us realise.
You know the feeling. You’re hunting for that one specific thing you know you read. A framework for prioritisation, maybe. Or that killer quote from a podcast guest. You’re sure it was in Lenny’s Newsletter.
So you open a tab. You search your email. You open another tab. You check your bookmarks. Another tab. Before you know it, you’re staring at ten browser tabs, a half-remembered thought, and you’re still no closer to finding what you need.
I’ve been there. Way too many times.
I’ve been a religious reader of Lenny Rachitsky’s newsletter for years. It’s probably the single most valuable collection of product management knowledge on the internet. But its value was trapped. Scattered across hundreds of Substack posts, YouTube videos, and Spotify episodes. The knowledge was there, but it wasn’t accessible.
Then, a little over a week ago, Lenny dropped a challenge.
He released his entire archive—350+ posts, 300+ podcast transcripts—as AI-friendly Markdown files. The challenge was simple: “build something.” The prize? A free one-year subscription. I saw the post, I looked at the calendar, and I thought… why not?
This is the story of how I took that challenge and built Lenny’s Product Brain, a full-blown Product Knowledge OS, in a little over a week.
From Frustration to a Plan
This wasn’t just about winning a subscription. It was about scratching an itch that had been bothering me for years. I didn’t just want to build a search bar. I wanted to build a system that understood how product people actually learn.
My thinking went something like this:
- What if you could just ask a question and get an answer grounded in the archive, not some generic ChatGPT fluff?
- What if you could see the connections between ideas? A visual map that shows you how, say, ‘growth loops’ and ‘retention’ are connected to what Bangaly Kaba talks about.
- What if you could get a structured learning path based on where you are in your career? A curriculum for your first 90 days, or for making the leap to product leadership.
- What if you could get personalised coaching? Advice tailored to your specific role, your company’s stage, and your biggest challenges.
Those four “what ifs” became the core of the app: Ask Lenny AI, a Knowledge Graph, Learning Paths, and a PM Coach. Plus a good old-fashioned Content Library to just browse everything.
The “Oh Sh*t” Moment
I built the app using React and TypeScript, but the real heart of it is the data pipeline and the Retrieval-Augmented Generation (RAG) architecture. Every newsletter and podcast transcript is chunked, indexed, tagged with topics, and linked to guests. When you ask a question, the app finds the most relevant chunks and feeds them to Google’s Gemini model to synthesise an answer, complete with citations.
But the build wasn’t all smooth sailing. The biggest roadblock was security.
This is Lenny’s life’s work. His IP. I had to make sure nobody could just scrape the entire archive. So I went into lockdown mode. I restricted database access, hardened the API, and even added CSS to prevent casual copy-pasting. I felt pretty good about it.
Then I pushed the changes. And everything broke.
The Knowledge Graph was empty. The stats in the Library showed zeros. Learning Paths wouldn’t load. It was that classic “oh sh*t” moment. In my quest to lock everything down, I’d locked out the app itself.
It was a good lesson. Security isn’t a switch; it’s a spectrum. The fix was to create controlled windows into the data—special functions that return just enough information for the UI to work, without exposing the raw content. It was a scramble, but it worked.
Designing for “Neural Warmth”
I didn’t want this to look like another boring SaaS tool. I wanted it to feel warm, inviting, and a little bit… brainy. I landed on a design language I call “Neural Warmth”—a dark mode aesthetic with amber and orange accents. The logo is a geometric brain, and that motif is woven throughout the app.
The Knowledge Graph itself was a blast to build. It uses D3.js to create a force-directed graph of all the topics, documents, and guests. It’s the kind of thing that sparks discovery. You can literally see how ideas cluster and connect.
What I Built, and Why It Matters
So, what’s live today at lennysproductbrain.vexto.app?
- Ask Lenny AI: Full RAG search with streaming responses and inline source citations.
- Knowledge Graph: An interactive visualisation of 508 topics, 439 documents, and 247 guests.
- Learning Paths: Curated content sequences for different PM career stages.
- PM Coach: Personalised coaching based on a 4-step onboarding process.
- Content Library: The full, searchable archive.
This isn’t about replacing Lenny’s newsletter. It’s about making the incredible archive he’s built more useful. It’s for the moments after you’ve read it, or for finding the right piece to read next.
This whole project started with a simple frustration and a LinkedIn post. It taught me that the gap between “content exists” and “knowledge is accessible” is bigger than most of us realise. And closing that gap, even for a single creator’s archive, is a crazy cool problem to solve.
If you’re a PM, I hope this tool saves you a few of those ten-tab-searching moments. And if you’re a builder, I hope this story is a reminder that sometimes the best projects come from just scratching your own itch.
Now, if you’ll excuse me, I have a link to post in a comment section.
Created by Mohammad Syed. A massive fan of Lenny’s newsletter and a hopeful contestant in his challenge.