On a Wednesday afternoon, I stared at a blank document for 20 minutes
I was juggling client projects, running my personal account, and trying to write a long-form piece. My material was scattered across chat logs, bookmarks, and memos, so every writing session began with half an hour of searching. I got stuck too: the more I saved, the less sure I felt about where to begin.
The method is simple: build a “work library”
Petered is an open-source tool for organizing personal knowledge. It saves webpages, books, podcasts, and other material in your own library, then uses AI to extract key points, connect them with older notes, and generate an editable writing draft. It doesn’t write an article from thin air; it turns “what I’ve read” into “what I can write now.”
I know Lin Ran, a friend who runs an independent podcast. She records two episodes each week in Hangzhou. She used to share only the episode links. Then she started putting interview outlines, listener questions, and her own post-listening reflections in one place. Six months later, she turned those notes into a short course called “Podcast Topics and Brand Partnerships,” using three cases from the archive to attract her first cohort. The tool didn’t write the course for her; it helped scattered material reveal its value again.
Try it today: 0 yuan, about 2 hours, copy and paste
When I wanted to test this, I resisted the urge to process everything at once. I set aside 0 yuan and roughly two hours. No programming was needed, but I did have to decide how my material should be organized. First step: open Petered’s project page, click “Try it” or “Get started,” create an account, make a new library, and drag in one article I read often. I didn’t start by importing thousands of items. I tested with 10: which content was worth keeping long term, and whether I could actually rediscover an old idea later.
This tool isn’t for everyone. If I already had a note-taking habit and could extract useful points manually, adding another tool might not be worth it. Not testing it yet is fine too. Putting the three client resources I use most in one folder can be just as effective a first step.
Match the setup to my stage—and avoid turning it into another project
Just getting started: If I could publish only once a week, I’d pick one fixed theme and save just 10 pieces. My goal would be to spend 30 minutes less searching next time, not to build a complete knowledge base.
With 1–2 clients: If I was already getting stable feedback from clients or a community, I’d save recurring questions, delivery records, and clients’ exact words separately. Each week, I’d turn one repeated question into public content, while giving potential clients a clearer sense of what I do well.
Scaling up: If several people were collaborating and content requests were increasing, I’d define naming and category rules before adding more sources or opening a team space. Once the operation grows, whether the team can understand and reuse the material matters more than how quickly AI can write.