About the bolt.diy category

Official - Open Source Community in partnership with StackBlitz (previously OttoDev) now BOLT.DIY

Detailed Guidelines:

  • Specifically for discussions around this project fork and its development
  • Should include feature requests, troubleshooting, and implementation discussions
  • Please post bug reports as issues on the GitHub repo
  • Please feel free to contribute through making pull requests in the GitHub repo
5 Likes

Hi there, could I get an executive summary on what oTToDev is and how it is different from the Bolt.new site?

I’m mostly curious about:

  1. If I set this up on my computer, will I run into limits that will require API keys or payments to LLMs?
  2. What should I expect in terms of speed compared to Bolt.new when my system is Windows, Ryzen 7 7800X3D, 64GB ram, GeForce RTX 4060 Ti?

Thanks, brand new to this but very excited!

1 Like

oTToDev is a fork of the open-source version of Bolt.new. Both projects allow you to develop, prompt, and deploy full-stack web apps using LLMs, but oTToDev is tailored to support custom setups.

oTToDev vs. Bolt.new:

  • Core Purpose: Both allow building, editing, and deploying full-stack web apps.
  • Difference: oTToDev supports local LLM hosting for cost-free usage with supported models, whereas Bolt.new relies on cloud-hosted AI.

Your Setup:

  • API Keys/Costs: oTToDev can eliminate costs by running LLMs locally.
  • Performance: Your high-performance system (Ryzen 7 7800X3D, 64GB RAM, RTX 4060 Ti) will leverage GPU acceleration, likely offering good speeds (speculation).

For more details:

2 Likes

Thanks @skpassegna really appreciate your reply.

I’ll definitely go ahead and give it a shot to install it on my machine.

I guess one benefit that the commercial bolt.new has is that you could access your projects from any web browser? Compared to the local install where you can only access it from your local machine?

Do we also know if it’s possible to move a project from bolt.new to oTToDev local?

I’ve just burned through 3 million tokens in a couple of hours using bolt.new commercial version. It’s absolutely amazing! It’s like I’m having a conversation with my very own CTO! Having some coding background and knowledge does come in handy!

It’s also BLAZING fast in the commercial version! I’m very curious to see the difference once I get the local version up and running.

I can totally see how as a new person you spend a lot more tokens while learning!

You’re absolutely right!

From my experience, the commercial version seems to have a well-tuned system prompt, which might be one reason it feels so fast. Regarding importing an existing project to oTToDev, I’m not entirely sure, but it might be possible in the future. When exporting a project from the commercial version, I noticed a .bolt folder inside it, which even includes a project-specific prompt. This suggests some potential for compatibility, but it’s hard to say for certain.

As for performance, how fast oTToDev runs will largely depend on your computer’s capabilities, especially if you’re using a local LLM. Interestingly, I suspect you might actually use fewer tokens with oTToDev while using remote providers.

It’d be great to hear from other experienced users who’ve tried both versions!

2 Likes

Hi, as a newbie, I’d like to know if anybody have any tips or advice as I’ll be downloading the code from GitHub and will like to do it as flawless as possible, as I’ve never done this before. :roll_eyes: Thanks!

2 Likes

Hi @missy ,
Welcome to the community.
To run oTToDev, just clone this repository on your local.

and to run the application, I would suggest to run it without docker. Follow this instructions: GitHub - coleam00/bolt.new-any-llm: Prompt, run, edit, and deploy full-stack web applications using any LLM you want!

if you face any issues or have any questions, please create a new Post and we will help you.

2 Likes

Thanks so much and sorry for posting in the wrong place. I now see how to start a thread here. Anyway, I took your advice and cloned it to my pc windows 10, all I need are the API’s and was wondering if the free API’s from glhf.chat can be used with oTToDev Bolt.new Fork? Thanks :wink:

1 Like

oTTodev does not support glhf.chat API as of now.
oTTodev supports, major LLM providers (Anthropic, OpenAI, Google, etc).

If you don’t want to use them and use free LLM, you can download LMStudio (https://lmstudio.ai/) or Ollama (https://ollama.com/). and you can use them for free and oTTodev has support for it.

Here is one of the video I found that addresses using oTTodev using local LLMs (https://youtu.be/8ommGcs_-VU?si=jLfV-sHqs6TFlaAd).

UPDATE:
@missy Is there any way to use glhf.chat api? - #3 by RustyR. you can try this.

3 Likes

keep it up ! amazing work and community

Hi,
I see you’re talking about this famous .bolt folder.
I tried to ask the commercial version’s chat what this folder was for, he told me it was a folder used by starblitz.
I think the prompt file is taken into account by starblitz before reading the user’s prompt to get that higher quality rendering that’s so specific to the commercial version, what do you think?

1 Like

Hello @missy welcome,
Sorry for my lateness. I was working on a clone I will make a pull request if accepted you could then use any openai compatible endpoints.

2 Likes

It only contains the prompt of the project. I have tried to add .bolt folder to some of my GitHub repo thinking that starblitz will allow editing with bolt but that did not work.
Yes, it is like the system prompt for bolt.new; it is prioritized.

1 Like

Fantastic! Keep me posted :grinning:

1 Like

Hey, I dont know if this is the right place to post this. I’m looking for someone who can help me develop a SaaS idea. Its a quite straight forward concept and should take to long. send me a dm on ig if you can help me @woonun

2 Likes

I’m a software developer with a passion for tech. I love where this is going and would like to contribute. One point of contention for your average user and getting it into their hands is always going to be setup, so I actually created a site that automatically updates from the main branch nightly, so people can just freely use it.

Maybe this can get more hands-on testing and using it, and of course I’d appreciate any feedback and don’t want to step on any toes.

For more details, please read:
No Setup Required! Batteries not Included! - General - oTTomator Community

I love this! How do I set up a version of OTTO dev that does have all of my API’s in it and can help me develop some ideas?

1 Like

Read the docs, but soon there will be a PR release to address this, basically a UI to manage your API keys from what I’ve seen. It’s pretty cool.

1 Like