Відео

I know it's a bit unclear in what state it is, PRs are not being merged so far. Should be community-driven somehow. Am pondering merging with his efforts and others there. I am planing a small video for tomorrow and will mention his there too. There is also Bolt.New office hours tomorrow, plan to ask there for their view and "plans" for the open source part of Bolt.

If you run npm run build and then npm run start it will show you hints PRess b for browser and it will open in browser

I did, contact, connected, looked into his repo, and considering to merge efforts There is also StackBlitz office hours tomorrow that I plan to visit and ask them about their views on open source part of Bolt

Sadly code is written to run on CloudFlare workers... And I am trying to run it on my digital ocean instance. Which proceed to be hard and I spent too much time on it so far. Can't decide between refactoring it to run on Node or figuring out how to run it with github.com/cloudflare/workerd I restarted the server and will make a video of how to run it locally tomorrow.

While I am figuring out how to deploy it on my own server in stable way I deployed to cloudflare You can try there as alternative 9356f8c3.bolt-2xk.pages.dev/

Thanks for the sub!

Will try to make a short video tomorrow

Thanks 💯

Yeah... I am suffering from that wrangler/cloudflare thing too... Spent most of my time around that so far and will need to spend more. I am running it in dev mode currently and it crashes 2-3 times a day. There is a way to run it in development mode locally And there is a way that was not added to run it in production mode too using github.com/cloudflare/workerd I can't yet decide if I want to refactor it to node.js or to try and use workerd Both have their pros and cons...

About running locally, will try to make a short video tomorrow

No, Bolt uses Web Containers that can do backend javascript. In theory python can be added too, there are web interpreters for it. There is another product called e2b.dev/ They have python in similar setup, but there it runs in containers on server side which is way more challenging thing to self host. Still, if you are interested in python on server side + open source, then take a look in that direction. It just works slower as its server side containers. And more expensive to run.

There is rate limit, it is still free though. I just rechecked and Open Router for free models gives 20 requests per minute and 200 per day, but this is per API key, you can add your own API key to get more and still use free models About 70%, you reference some kind of comparison for Gemini 1.5 Flash?

@@louisfeges2913 thanks for kind words Luis. Trying to strive for what you describe.

Do you want to use it with local models too? There are some forks that already support that May be I will do a short video later on that

which mode is selected for you? gemini?

@@EduardsRuzga yes. Gemini 1.5

@@nocapnobs7845 yeah, getting it too, will be investigating this week

While I am figuring out how to deploy it on my own server in stable way I deployed to cloudflare You can try there as alternative 9356f8c3.bolt-2xk.pages.dev/

Lol, that's too much praise :D So many before me in StackBlitz, OpenRouter, Google and on and on Thanks! Still things to do for this to be more useful.

What do you like about bolt.new-any-llm? I would love for the community to sync on 1-2 forks, currently, there are a dozen forks where some work is being done. Am exploring whether it makes sense to merge with bolt.new-any-llm Not sure yet as there are two "directions" with this that are bit hard to synchronize currently. One is using it with local things like docker, ollama, lm studio. Which is not a good thing for total nondevelopers. Another one I am interested in is a hosted solution on websites that require setup to run. And I am not yet sure how to allow a single project to serve both directions.

seems I have some issues, second crash of the day, give me a sec

Restarted, will try to investigate bit later of why it crashes

@@EduardsRuzga Fix it please broooo

What do you mean by attachments? Images?

@@EduardsRuzga yes, images. In Bolt chat i do not see the option to upload an image.

@@HonorHasCome it is possible to add, not on top of my list yet, with multiple model selection part of the problem is that not all models support it But noted, will see bit later

@@EduardsRuzga tnx

You mean local llms running on your machine? My focus is on hosted solutions that do not require local setup. I wonder if one can use a hosted solution that connects to localhost as a provider. I use LM Studio myself never found local models good enough

Preview seems not to work due to problems with running terminal commands, it's on my list of things to fix next As for AWS bedrock models, no, I saw some others forks adding that, check forks if you need it in timely manner I hope we can organize in to community that adds such things together, plug and play style. Need AWS models, just flip a switch

RAG of what type? RAG is basically indexed text snippets to be added to context when needed, what should be indexed and pulled out?

There should not be, but I just had a crash of my server, will need later to investigated what happened

@@EduardsRuzga thank you very much, please do update us when the investigation is complete

While I am figuring out how to deploy it on my own server in stable way I deployed to cloudflare You can try there as alternative 9356f8c3.bolt-2xk.pages.dev/

Thanks!

Спасибо!

the link to yours isnt working would love to work on this with you further keep up the great work

@@ChrisMateo-q6s there is other popular repo, i am thinking weather its worth merging efforts or not considering there are interests of running it it locally and in self hosted variant.

While I am figuring out how to deploy it on my own server in stable way I deployed to cloudflare You can try there as alternative 9356f8c3.bolt-2xk.pages.dev/

No, in my fork its open router now. Upload can work but i need to try. It msy glitch in unexpected ways. I think I can add antropic back as option. Probabky need to make popup setting fir choosing provider and model, too much for header.

@@EduardsRuzga thanks!!!!

@@Ony_mods hey, yeah technical mistake on my side. I will do recap video later that will be shorter and rerecord audio better. Will leave comment here when its published.

I published video a day ago ua-cam.com/video/wAjewqnzq7M/v-deo.html

Did you use any others? Replit and GetLazy feel even worse for me. What I am excited about here is that it has open source, self host variant. I already extended it to use free models, no need to pay. Need to host it now. Soon ;)

Yeah, sadly being in eastern EU and having kids home in the second half of the day puts me not in the best spot for live streams for the Western hemisphere.

@@EduardsRuzga family first, keep up the good work

@@jpoole4931 agreed and thanks!

I agree, am actually working towards having a platform that allows ChatGPT to run code in the browser. But there is more than that and it's not the main part of it.

@@WaveOfDestiny yeah, could be. I am heavy user, use it for writing, brainstorming, summarising, making image ideas. But there are so many free tools out there that if you are occasional user you can do with using free tiers of a dozen of offerings.

This was first impressions video and here it did quite well. I am testing it more and there are issues. It does fail with complex files. I also tried one workflow that did not work for me anywhere else. Give it a UA-cam transcript in a canvas, ask to think of good places to split it to chapters in chat, then add them in appropriate places of transcript as comments. If you ask to do the mall at once it fails, but one by one it succeeds and Claude or Gemini can't do it well... And one can imagine similar tasks for code. You can ask it to make comments, and then selectively apply them. That is different from Claude that does everything in one shot and also fails spectacularly depending on complexity of the task. I really hate how Claude always rewrites and as task becomes more complex it rewrites with more and more mistakes... Will see over next weeks as I find more of productive workflows with it Anyways, that is my experience, yours is different, care to share more? What tools do you use?

@@tomascoox yeah, tried it once so far. I would compare it to gptengineer, replit agent, getlazy though. It is solely focused on coding and has deploy options. First impression of bolt was good, better then alternatives. What is yours?

I use Voice In It's not superb, multiple other solutions are not good too chromewebstore.google.com/detail/voice-in-speech-to-text-d/pjnefijmagpdjfhhkpljicbbpicelgko The best one is in ChatGPT mobile and desktop apps. But that can not be used out of them. I am thinking of making my own extension that uses webgpu accelerated whisper Based on this huggingface.co/spaces/Xenova/realtime-whisper-webgpu

Not only AI, my wife thinks so too :D And she is not AI :D But in all seriousness, yes, that is one area to improve.