I like how loveable.dev understands the requirements and tell us in more structured way what it understands. but it is expensive. I am looking for a bit cheaper or opensource alternates.
Welcome! Bolt.diy can be completely free using Gemini Flash 2.0, HuggingFace, etc. You’ll just need to try it out and see if it meets your needs. Generally, you pay for convenience.
Yes, it can but you need to now, that bolt.diy is…
- still in early development
- not optimized to one LLM/Provider as the commercial loveable or bolt.new, which means, you have to put some more effort in doing good prompts yourself. Use the “Enhance Prompt”, as well as use Starter Templates go get better results.
- Not every models you choose performs best. So stick at beginning at least to the mentioned ones in the docs, to have a chance to get something good out.
- Take a look at my youtube videos for uncut/live showcase of some bolt.diy (I mention this, cause there a lot of youtubers dont show the actual state and just make perfect videos with prepared stuff, what leads into a lot of newby users, which think everything should work perfect and it just does not for them, whats not the case)
when running bolt.diy from your computer does it slow everything down or cause your system to lag at all? is their system requirements to know about before running it locally?
It depends. I’d say at least 1 CPU and 4GB RAM “Available” (meaning dedicated and not used by anything else). So if you have a multi-CPU with at least 8GB of RAM, you should be fine.
If you are running full blown “dev” mode by cloning the repo and running it, you are running a Remix server and other things. On top of that, each command that the AI runs is running locally, so like downloading dependencies, executing code, etc. All of that, in conjunction will internet calls (API, Web Container, etc.) hit your network resources and some hard drive IO.
So basically if you have say a low end Chromebook or something, or low memory (also running in a Docker instance on Mac or Windows) you will likely feel it yes. Docker on Windows and Mac are virtualized (native only to Linux), causing overhead, and at least on Mac the default is limited to 4GB RAM. But it’s kind of hard to say, because it really depends.
If Bolt.diy had a direct connection to supabase, I know where you would find a lot of users - L0vable is in free fall right now. I actually believe some users are buying pitchforks. : )
You could also try jdoodle.ai. Fairly new but easy to use and efficient!
i didnt even notice your response… thank you… im totla beginner and totally discouraged now… i got it running and i couldnt even type a sentence in another window without having to wait 15 mins for it to show up on the screen… everything lagging hard …
Hey @SharkSavvy why do you say lovable is in a free fall right now?