Founder Recap Dec 2025
Founder Recap Dec 2025
A quiet day of reflection and looking back on all the things that have happened since the beginning of the journey.
My goal for releasing the public repo for Rekait by Christmas failed, followed by failing to release the repo before New Years. Sorry to the few people that were actually looking forward to try it, I shouldn’t make more promises but maybe the third time’s the charm? January public repo release, let’s go!
I just want one very simple end to end experience. Just one integration. Discord. As long as you can talk to the Rekait agent to build a custom pipeline that feeds into a discord bot, I’m happy. That’s all I want for the first release and I’ll reach out to early users that I know want specific integrations e.g. MongoDB, LINE Webhook, etc. We can go step by step from there or maybe the Rekait agent is good enough to help you out.
The demo was easy to build fast and I don’t mind presenting a janky demo, but something about an open-source repo, where someone else will run my code, makes me feel icky if I ship something full of bugs. It also doesn’t help that I had already cut a lot of corners shipping the demo. The whole point of Rekait is to have a solid foundation to iterate upon and if these foundations are shaky, everyone’s houses are going to fall apart even if it’s a simple hut. Maybe I’m walking down the path of over-polishing things, but I feel like I’ve set a low bar already, ONE integration and that’s it. Low bar doesn’t mean full of bugs right?
There was an option of just building the discord integration into the demo but that thing is so fragile, I hate the idea of building on it further and there are features being blocked by that design that would make me do work arounds even more. It’s also somewhat whacky that the demo actually uses express.js as the main backend server which spins up a separate python process instead because that was fastest path for demo and it kind of snowballed downhill from there.
Maybe I might look back at this year and feel dumb for moving so slow compared to all the other startups out there. Feels like I shouldn’t compare too much but I also don’t want to give myself excuses if I really am moving too slow.
Trying to do an open-source startup feels so hard. I know I can’t just sit and think up all the possible scenarios and sometimes it’s best to tread forward to let the path reveal itself. And I do feel that every passing day, every interaction, the idea crystalizes more and more compared to when I started with a vague sense of the problem I wanted to solve, which stems from my personal frustrations and pains in managing AI backends.
The two biggest struggles for me this past six months was basically two things: who are my customers and what is my pricing model. Both of which I may have added complications to when I decided to open-source things. Maybe I should just not worry too much and keep building, but I can’t help to just sit, write, and ponder about customer and pricing because it feels so scary if I can’t give a confident answer to the question of why would someone pay for my product, even if it doesn’t fully exists yet. But there’s only so much you can guess in your head, so I really need to get a usable enough product for someone to try and see whether they actually would pay.
I have a better sense now of what the initial version for pricing would look like, the back of the envelope math checks out but actual margins and cost remains to be seen. Reality might be different after implementing it and having live users, let’s see if all the pieces fit.
North star for next year:
How will you serve the world? What do they need that your talent can provide? That’s all you have to figure out.
– Jim Carrey
Screencap, the UI still looks pretty much the same although interactivity is smoother. More importantly, under the hood is a very different beast. 