@heyandras It would be worse if you haven't done this project. No value created for others and your team, no learning.
English
Gabor Barna
350 posts
Currently in the worst case scenario, but there are so much things to do. 😅
That one person: 💜
Andras Bacsai@heyandras
my plan with shoutrrr.com is to make around $5k/m with it, so i can hire another person to make new apps. not sure if we can make it, the market is full of social schedulers. but worst case scenario is that people will get a free and opensource scheduler, not bad.
English
@SpirkovskiZoran @heyandras @Jean_build Tailscale and web interface works perfectly, and you can enable Jean MCP and connect it to Hermes, Openclaw and other harnesses.
English
@heyandras @Jean_build I need to give it a shot. Just set up my own workspace on a server. I haven't given it much consideration so far. Does it have anything like messaging to apps like Telegram, or do you just SSH in to run it?
English
Since I switched on only server based dev with @jean_build , I could build all day without a power cable. I was missing this a lot.
English
@daaysorn @heyandras @Jean_build I think it is more dependent on the project, doing local validations, builds or having parallel test deployments locally.
English
English
@kirusfg1 @dr__madmax @thdxr How did you get it working?
I only had issues with hotkeys, sound device, language and type back. Everything else (eg. the UI) is working fine. On Ubuntu, so not even compiling the system and such.
Thanks for describing the tricks!
English
@simonlehmannn @heyandras Maybe, I haven't checked what the AI companies offer nowadays, and sure you can use ssh/mosh, tmux and such to remote work. But then you are in vendor lock in. Jean is vendor agnostic and on top of that offers an opinionated, configurable workflow where you can work in parallel.
English
@robaganrab @heyandras But that one I can do with Codex/Claude Code as well?
English
@simonlehmannn @heyandras You have an idea on the go, by the time you get home, Jean implemented it or even deployed it self hosted.
Corporate AI backend (eg. free tokens), Jean can work in one or many repos, investigate, cross reference, even while my team have lunch and comes up ideas to look into.
English
@heyandras I was about to use Jean lile 5 timea, but I still dont get the use case. As an indie developer with multiple SaaS, whats are the pros?
English
@lepointcrypto @heyandras Run Jean on WSL, and don't bother with the UI with the extreme cursor size. Have a remote machine at work and drop ideas to be implemented while out to have lunch with colleagues. Have a machine at home, running Openclaw or Hermes and interacting with Jean MCP.
English
@heyandras Never really understood the use case of headless tbh. What’s the use case ?
English
Jean could work on remote servers, someone also made it dockerized, so it is easy to deploy anywhere.
So maybe I just need to improve this flow? Like be able to switch between remote instances.
Theo - t3.gg@theo
I’d estimate we’re ~6 months from most devs moving their code agents off of their laptops
English
@heyandras Jfrog Artifactory deprecated a type of tokens a few years ago and promised that the new token type will enable fine grained permissions. Till the current date tokens are still all in equal to passwords, just instance specific.
English
Hit me with the best fine grained permission system that you ever see in any app.
GIF
English
@graasdev @heyandras @grok I had a branch for Gitlab locally trying to support the respective Github features (issue board, PR/MR, etc.), but I'm not sure how and if to contribute. I can definitely test the features. What do you say @heyandras, shall we open PR for Gitlab support?
English
🚀Shipping Jean v0.1.59.
It has full @grok Build support and with lots of fixes, improvements.
GIF
English
@antideologist @GergelyOrosz Or companies can have a private "GPU" farm and developers can use the loaded models.
English
@GergelyOrosz "local"
=> meaning companies will need workstations
=> will need people to manage these workstations
=> they will have running costs
=> laptop run models will conflict with other software / eat the RAM and get computers to heat up
English
There is massive irony in how AI coding tools are starting to become TOO expensive for many enterprises - after eg Anthropic removed subsidizing AI subscriptions.
We might go from "everyone use AI for everything!" to "you have $300/month AI budget; use your brain for the rest."
English
@heyandras If one or more higher layer depend on your product, your will become infrastructure and therefore it gets harder to change. Distant users' workflows will depend on you. For example people don't really touch electric wires anymore. 😉
English
Just read Simon's latest newsletter and it got me thinking...
I have an idea which is perfectly fits this idea, so I will execute after I clean Coolify / Jean related stuffs this week.
What do you think?
Can we build a headless SaaS?
English
@AdityaTripathiD @heyandras @Jean_build I only tried the GUI version, but I've seen the headless option. You might have to steer the agents to SSH into the WSL VM for now. Jean itself is not a resource hungry software, so you can work this way.
English
@robaganrab @heyandras @Jean_build Ah Jean headless in Linux? I meant being able to use the Windows version but with WSL instead of installing the Linux version
English
Be resilient, use @Jean_build.
If one provider enshittificated, you just click and you can continue yoloing on another provider. 🤘
ps: I will release a new version tomorrow with huge improvements on Codex And Opencode backends.
Nico@nico_jeannen
Idk what's going on with @AnthropicAI but they messed up the weekly limits badly it seems Anyone else with the issue? I've almost never hit the weekly limit on the x200 plan (except maybe 2-3x a few hours before the reset) and now I'm hitting it on Monday lol
English
@AdityaTripathiD @heyandras @Jean_build I use Jean on WSL. Web access over Tailscale so the cursor is not made for grandma. Otherwise it works nicely.
English
@heyandras Could it be somewhat discounted for you?
x.com/i/status/20202…
Boris Cherny@bcherny
Excited to hear what you think! Pro & Max users have $50 in free credit to try it out, and it’s 50% off until Feb 16.
English
Anthropic has significantly improved their costs, as the same number of tokens with Opus 4.5 versus Opus 4.6 is much lower.
Opus 4.5 -> 157M -> $126
Opus 4.6 -> 155M -> $6.95
(this is not a scientific conclusion for sure, but yeah)
English
🛠️ Merge local worktree's to base branch works.
If you don't want to open a PR for each worktrees, you can do it locally.
With this way, your git history will be preserved while still having the benefits of isolated Git worktrees.
English
@heyandras Add a scroll stopping and a few lines from the next file. If the user scrolls again, it can switch to the next file. This gives the program enough time to pull and render the next diff. This can work upwards as well.
English
🛠️ Refactored the git diff view, because who wants to review all the files in one long list?
English
@heyandras Name every project of yours coolify-*, like coolify-jean. Problem solved! 😅
English
everytime i work on something new and exciting, i always feel bad that i am not working on Coolify.
how to overcome this?
English
@heyandras Now is the best time to get a puppy! Additionally you can check every 2 to 3 hours if there is internet already, even during the night. 😅
English
Week 2 without internet at our new home.
(ISP not able to build out the cables)
How is your Friday? 😅
English
