Why does this reference Lovable at all? The tagline for this project is "Clone and recreate any website as a modern React app in seconds" and requires a FIRECRAWL_API_KEY in order to perform the necessary scraping.
Lovable is more of a self-contained LLM chat which creates a new react site with optional supabase integration.
I still sometimes have to refer to OpenHands as "formerly known as OpenDevin". It may be limiting in the long run, but it definitely had some short-term punchiness.
It's not really an "open lovable". I mean, there is of course a NextJS repo as part of it, but you need to bring your own API keys to an AI provider, as well as e2b.dev (sandbox, required) as well as firecrawl.dev (web scraping, oddly required).
It's the web scraping more than anything that makes me the least enthusiastic. I get that's the core of their business, but packaged this way, it's much less a loveable-style "build your vision" and more a "quickly copy another site". Just serving as a copy machine is the least interesting use of all this insane compute.
It feels like a common thing in the AI space (moreso than other OSS spaces). People don’t get how trademarks work and that you can’t just name your project after something that’s already popular. I think some people think it’s a clever growth hack.
nope, fire your lawyer. you get a common law trademark by trading under a name. no need to register it. TM registration is just to streamline big corps chasing after lots of small abusers in a wackamole way, that’s all. if the abuse in question is as egregious as “I used their product name, barely modified, to market my own thing, which is not just in the same general market but a literal clone of their product” then a judge is not going to say “Ah but I see they didn’t send in a form to register their trademark, so yeah carry on with blatantly stealing their work lol”
I feel like "also trademark registration is not remotely relevant here" is a distracting statement then. Not necessarily wrong but easy for people to bikeshed over.
Can we run an agent in Github CI? I create a Github issue with my requirements, the CI runner runs Aider scripting mode (it connects to say, OpenAI api) which looks at the comments, maintains a branch, runs tests, takes browser screenshots and then gives its replies.
Why would I do this? It would be better to have an app that turns a react app into a native web app. Speculation rules and page transitions make it possible already.
Reminds of me of tools like https://huu.la, the main difference is newer generations focuses on utilizing LLM to map design to code, versus the older generations focus on using WYSWYG tooling to edit the page directly.
Why does this reference Lovable at all? The tagline for this project is "Clone and recreate any website as a modern React app in seconds" and requires a FIRECRAWL_API_KEY in order to perform the necessary scraping.
Lovable is more of a self-contained LLM chat which creates a new react site with optional supabase integration.
Firecrawl is a highly dubious company. Their job ads for bot "agents" seemed like a try to get free labour.
Wait is this a Lovable clone that is actually called open-lovable? That seems bold
Having started OpenDevin (now OpenHands [1]) I can say it's definitely worth renaming. It's very limiting attaching your branding to someone else's
[1] https://github.com/All-Hands-AI/OpenHands
idk, do you think you would have gotten the start you did NOT being called opendevin?
(hi btw! see GP's talk on OpenHands here https://www.youtube.com/watch?v=o_hhkJtlbSs&t=4s)
I still sometimes have to refer to OpenHands as "formerly known as OpenDevin". It may be limiting in the long run, but it definitely had some short-term punchiness.
It gets you some quick attention.
The people that need to notice will follow.
It's not really an "open lovable". I mean, there is of course a NextJS repo as part of it, but you need to bring your own API keys to an AI provider, as well as e2b.dev (sandbox, required) as well as firecrawl.dev (web scraping, oddly required).
It's the web scraping more than anything that makes me the least enthusiastic. I get that's the core of their business, but packaged this way, it's much less a loveable-style "build your vision" and more a "quickly copy another site". Just serving as a copy machine is the least interesting use of all this insane compute.
It feels like a common thing in the AI space (moreso than other OSS spaces). People don’t get how trademarks work and that you can’t just name your project after something that’s already popular. I think some people think it’s a clever growth hack.
As long as it’s not misleading: https://docs.github.com/en/site-policy/content-removal-polic...
I am reminded of OpenNext (https://opennext.js.org/) even though Vercel has a Next.js trademark.
compliance with a github policy isn't really the issue lol
Im waiting for open-open-ai
Everyone's waiting for open-open-ai.
You're in luck! https://openai.com/open-models/
Hardly.
With “open” being such a common word, I wonder if that one might actually get a pass. And if not, then maybe open-squared-ai.
It seems like trademark infringement.
Can’t find a Loveable trademark for the app AI
Well, for starters, it's not spelled with that first e.
https://tsdr.uspto.gov/#caseNumber=99194755&caseSearchType=U...
also trademark registration is not remotely relevant here
What do you mean? It's literally the legal mechanism by which Lovable Labs Incorporated can prevent this guy from calling his thing open-lovable.
Registering your trademark makes enforcement easier, but there's a reason both TM and R exist - unregistered trademarks still hold power.
Sure. But for a company with a lot of money at stake, filing a trademark application isn't a huge bar and makes winning the legal case a lot easier.
nope, fire your lawyer. you get a common law trademark by trading under a name. no need to register it. TM registration is just to streamline big corps chasing after lots of small abusers in a wackamole way, that’s all. if the abuse in question is as egregious as “I used their product name, barely modified, to market my own thing, which is not just in the same general market but a literal clone of their product” then a judge is not going to say “Ah but I see they didn’t send in a form to register their trademark, so yeah carry on with blatantly stealing their work lol”
I feel like "also trademark registration is not remotely relevant here" is a distracting statement then. Not necessarily wrong but easy for people to bikeshed over.
fair I retract “is not remotely relevant here” and offer “is not really relevant here”
Lovable is a $1.8 B valuation company with $75 M ARR. Using some thousands of dollars to defend that stake sounds reasonable.
I didn't know about https://e2b.dev/ but I was looking for something exactly like that. Does anyone know about any self hostable alternatives?
Can we run an agent in Github CI? I create a Github issue with my requirements, the CI runner runs Aider scripting mode (it connects to say, OpenAI api) which looks at the comments, maintains a branch, runs tests, takes browser screenshots and then gives its replies.
https://aider.chat/docs/scripting.html
It appears that e2b runs Firecracker microVMs (https://e2b.dev/blog/how-manus-uses-e2b-to-provide-agents-wi...)
It shouldn't be too hard to get a Firecracker orchestrator running locally - the articles here were very helpful when I was doing this myself: https://jvns.ca/blog/2021/01/23/firecracker--start-a-vm-in-l...
You should look at [Modal](https://modal.com/), not affiliated.
Lovable runs on Modal Sandboxes.
Played with https://www.freestyle.sh this weekend, really solid
you can host e2b: https://github.com/e2b-dev/infra/blob/main/self-host.md
we host it for https://www.definite.app/. You'd need pretty heavy usage to beat e2b's pricing.
You can check out https://daytonaio-ai.framer.website/, it’s also could be self-hosted
Great initiative! I'd love to see this project build a community around it. Quick question, what do you need Firecrawl for?
Why would I do this? It would be better to have an app that turns a react app into a native web app. Speculation rules and page transitions make it possible already.
Reminds of me of tools like https://huu.la, the main difference is newer generations focuses on utilizing LLM to map design to code, versus the older generations focus on using WYSWYG tooling to edit the page directly.
Can I run the whole chain FOSS? Firecrawl, this (maybe the LLM can be local too?)
Now… any chances we could get a thing that takes in a React app and spits out a normal website with minimal or no JS?
Yes, it’s called `tsc` ;-)
How would you know it works if you can't add an imperial megaton of tracking scripts? /s
There's also bolt.diy which uses browserbased webcontainer to build apps.
That’s not open though right. Webcontainer is closed
Good point. codesandbox has an equivalent thing - i guess that could be a replacable thing.
Is this a desktop app wrapper for lovable?
It needs a Docker-Compose.
Why not use bolt? It's open source and is competitive with lovable and replit.