A wiki you don’t interact with is useless.
I don't want my "second brain" or wiki to just sit there.
So I built wikikarp, a LLM wiki viewer. The goal is to turn your personal wiki into something you can actually use every day.
Let's get to level 20 and evolve! 👉 wikikarp.app/early
#buildinpublic
Sure!
Probably learn as much as possible about the problem space and what user wants.
#buildinpublic has always been a fun place to share the journey and discover cool projects. And TradingGrader is pretty cool! How about you?
Thanks! Been building model0.app for structured prompt management and only this week I’ve extended it to help developers run agents all night long.
Finally got in!
#buildinpublic
Running agents all night without blowing through your quota is tricky.
So we're working on it @ model0.app
#buildinpublic
Very cool tracker! It also supports data import! I can import agent data to track its productivity to see which agent is most productive lol
Part 2 of the demo
I have too many side projects and got tired of babysitting AI agents.
So I built a way in model0.app to assign tasks straight from TODO.md, and just let them run to completion.
#buildinpublic
It's coming together...
#buildinpublic
Agents deserve a name. It was hard managing agents cognitively without a name.
Building new agent management feature @ model0.app
#buildinpublic
Too many agents to manage?
We're extending our solution & building it @ model0(dot)app
#buildinpublic
Do we actually need more agents?
Feels like everyone wants to "agentmaxxx", run it while you sleep, run, wash dishes, or simply breathe... Everyone wants to do more. Sometimes it's counter-productive. I guess the question should be how many agents can we manage? How should we manage them?
Improved structured output schema validation for Claude models in model0.app!
Alpha is out 👉 model0.app/alpha
#buildinpublic
quality-of-life improvements
#buildinpublic
Alpha for model0(dot)app is live 🚀
Still rough around the edges. Join the waitlist if you want early access + want to help us break things 🚧
#buildinpublic
Getting LLM output schemas right is surprisingly painful 🤕
Added built-in schema validation to model0(dot)app so I don’t have to guess anymore
#buildinpublic
I never felt good seeing apps storing API keys in localStorage/IndexedDB.
So I use Keychain to encrypt them in model0(dot)app 🔒
#buildinpublic
Finally got automatic updates working on model0.app 😅
Took longer than expected, but now we can always bring you the latest feature/security updates to our prompt management tool!
#buildinpublic
I mess up output schema every single time 😩 So I've added a sample output schema whenever you create a new prompt in model0(dot)app.
#buildinpublic
Ever needed to run an older prompt in prod… but can’t find it?
Now you’re digging through git history 😅
What if you could just save prompt versions and run them locally with any model - no scripting required?
We just added prompt versioning to model0!
#buildinpublic
Sometimes you're just "one" prompt away from fixing prod.
Finding that prompt becomes critical - so model0.app now has CMD+K search 🎉
I've built model0.app because experienced developers want to encode taste into prompts, not rely on generic ones 😴
Check out our waitlist → model0.app
#buildinpublic
Some people just wait, I refresh everyday🧘♂️ Excited to be waiting for more processing and waiting! One more step closer to making prompting DX more delightful!
#buildinpublic
model0.app now supports more LLMs 🚀
Run the same prompt with different datasets across models — and evaluate structured JSON outputs in seconds!
#BuildInPublic
bsky has a limit of 3 minutes per video. Here's the second part of the demo!
Want to test your prompt template against different datasets?
model0.app now lets you inject variable sets into your prompts. Ad hoc output evals in seconds!
Interested? Join our waitlist → model0.app
#buildinpublic
100%! Definitely will improve the error messaging along the way with as much details as possible
Good question! If the Markdown/CSV syntax is invalid, we fall back to showing the raw JSON. If the JSON is invalid, we display an error message.
