[Frontier] Flow: Your Brainstorming Copilot
a product launch 🚀
This is a newsletter on sharing my journey to building something wonderful. To receive Frontier in your inbox 1~2 times a month, subscribe here:
Launching Flow 🌊
I talked briefly about Flow in my previous newsletter. I’m happy to announce we’re launching Flow on Product Hunt today 🥳
Flow is a visual brainstorming tool that brings markdown and AI to a canvas. You can place Notion-style text blocks by clicking, connect them by dragging. Think of it as a 2D version of Notion.
We’re also bringing exciting AI features to Flow. Upon connecting two blocks, a textbox and “generate” button is presented. Fill it in with whatever you want AI to do (e.g. summarize, ask a question) and click the generate button. It will fill the target block with AI generated text.
Check Flow out on ProductHunt!
It would be much appreciated for readers who have Product Hunt accounts to upvote Flow 🙌
A Few Thoughts Behind the Scenes.
Flow was built in 2 weeks, 100% bootstrapped.
The idea came to me back in the summer, when I was writing reports on Pinpoint. I usually brainstormed my ideas before I started writing. This was done on Figma, as other tools did not offer a matched level of flexibility. Since Figma was a design tool, I could do whatever I wanted with it.
There were a few problems:
No markdown support
Constant switching tabs to ask Google / GPT just to copy paste onto Figma
Information became disorganized as I expanded my mindmap
It took a month just to distill the information from the canvas to a fully-fledged report
I guess I’d start building a tool that could solve my problems with Junseon, a fellow writer on Pinpoint who had been going through similar headaches. We started building in the middle of February, finished at the beginning of March.
I had no prior experience in web development + Junseon didn’t have much knowledge in BE + schema design. And it took us just 2 weeks to build Flow. Coding has become 3x easier due to Copilot and GPT4. So, as a side note for aspiring builders: don’t talk, just start building! (Granted, we did receive help from fellow builders. Huge thanks to Gilmo and Jaehong for saving our asses ❤️)
More on the launch:
The reason for the premature launch is: perfecting Flow in its current form is pretty meaningless. The market of “AI productivity tools” is filled with competition, and if we don’t come out 10x better, there is a very low chance of winning. For one, there already is an amazing tool called Heptabase (backed by YC and Kleiner Perkins, #1 award from Product Hunt), with a similar premise, built for over 3 years 🙃
Even if we made Flow 5~10% better, it would be something like 0.25x Heptabase. So we’re going out to the world, launching quickly, as soon as we had a quantum of utility. In hindsight, this could be a mistake. But experiencing the cycle of building and launching globally is very humbling, and many lessons are being learned.
Building a global audience is something that I’ve never done before. Identifying watering holes, building GTM, promoting on SNS… all these things need to be done in English and on a budget close to zero. I realized this was going to take way more time and effort than I expected. But you already know the story: there’s no choice other than trying hard again and again, improving on every iteration.
Thank you for reading this issue of Frontier by Doyeob.
Hope you enjoyed this new style of writing.