This fork of Bolt.new (oTToDev) allows you to choose the LLM that you use for each prompt! Currently, you can use OpenAI, Anthropic, Ollama, OpenRouter, Gemini, LMStudio, Mistral, xAI, HuggingFace, DeepSeek, or Groq models - and it is easily extended to use any other model supported by the Vercel AI SDK! See the instructions below for running this locally and extending it to include more models.
Bolt.new is an AI-powered web development agent that allows you to prompt, run, edit, and deploy full-stack applications directly from your browser—no local setup required. If you're here to build your own AI-powered web dev agent using the Bolt open source codebase, [click here to get started!](./CONTRIBUTING.md)
- **Full-Stack in the Browser**: Bolt.new integrates cutting-edge AI models with an in-browser development environment powered by **StackBlitz’s WebContainers**. This allows you to:
- Install and run npm tools and libraries (like Vite, Next.js, and more)
- **AI with Environment Control**: Unlike traditional dev environments where the AI can only assist in code generation, Bolt.new gives AI models **complete control** over the entire environment including the filesystem, node server, package manager, terminal, and browser console. This empowers AI agents to handle the whole app lifecycle—from creation to deployment.
Many of you are new users to installing software from Github. If you have any installation troubles reach out and submit an "issue" using the links above, or feel free to enhance this documentation by forking, editing the instructions, and doing a pull request.
2. Install Node.js from https://nodejs.org/en/download/
Pay attention to the installer notes after completion.
On all operating systems, the path to Node.js should automatically be added to your system path. But you can check your path if you want to be sure. On Windows, you can search for "edit the system environment variables" in your system, select "Environment Variables..." once you are in the system properties, and then check for a path to Node in your "Path" system variable. On a Mac or Linux machine, it will tell you to check if /usr/local/bin is in your $PATH. To determine if usr/local/bin is included in $PATHopen your Terminal and run:
3. Rename .env.example to .env.local and add your LLM API keys. You will find this file on a Mac at "[your name]/bold.new-any-llm/.env.example". For Windows and Linux the path will be similar.
If you can't see the file indicated above, its likely you can't view hidden files. On Mac, open a Terminal window and enter this command below. On Windows, you will see the hidden files option in File Explorer Settings. A quick Google search will help you if you are stuck here.
To make new LLMs available to use in this version of Bolt.new, head on over to `app/utils/constants.ts` and find the constant MODEL_LIST. Each element in this array is an object that has the model ID for the name (get this from the provider's API documentation), a label for the frontend model dropdown, and the provider.
By default, Anthropic, OpenAI, Groq, and Ollama are implemented as providers, but the YouTube video for this repo covers how to extend this to work with more providers if you wish!
When you add a new model to the MODEL_LIST array, it will immediately be available to use when you run the app locally or reload it. For Ollama models, make sure you have the model installed already before trying to use it here!
-`pnpm run start`: Runs the built application locally using Wrangler Pages. This script uses `bindings.sh` to set up necessary bindings so you don't have to duplicate environment variables.
-`pnpm run preview`: Builds the project and then starts it locally, useful for testing the production build. Note, HTTP streaming currently doesn't work as expected with `wrangler pages dev`.
-`pnpm test`: Runs the test suite using Vitest.
-`pnpm run typecheck`: Runs TypeScript type checking.
-`pnpm run typegen`: Generates TypeScript types using Wrangler.
-`pnpm run deploy`: Builds the project and deploys it to Cloudflare Pages.
This will start the Remix Vite development server. You will need Google Chrome Canary to run this locally if you use Chrome! It's an easy install and a good browser for web development anyway.
- **Be specific about your stack**: If you want to use specific frameworks or libraries (like Astro, Tailwind, ShadCN, or any other popular JavaScript framework), mention them in your initial prompt to ensure Bolt scaffolds the project accordingly.
- **Use the enhance prompt icon**: Before sending your prompt, try clicking the 'enhance' icon to have the AI model help you refine your prompt, then edit the results before submitting.
- **Scaffold the basics first, then add features**: Make sure the basic structure of your application is in place before diving into more advanced functionality. This helps oTToDev understand the foundation of your project and ensure everything is wired up right before building out more advanced functionality.
- **Batch simple instructions**: Save time by combining simple instructions into one message. For example, you can ask oTToDev to change the color scheme, add mobile responsiveness, and restart the dev server, all in one go saving you time and reducing API credit consumption significantly.
### How do I contribute to oTToDev?
[Please check out our dedicated page for contributing to oTToDev here!](CONTRIBUTING.md)
### Do you plan on merging oTToDev back into the official Bolt.new repo?
More news coming on this coming early next month - stay tuned!
### Why are there so many open issues/pull requests?
oTToDev was started simply to showcase how to edit an open source project and to do something cool with local LLMs on my (@ColeMedin) YouTube channel! However, it quickly
grew into a massive community project that I am working hard to keep up with the demand of by forming a team of maintainers and getting as many people involved as I can.
That effort is going well and all of our maintainers are ABSOLUTE rockstars, but it still takes time to organize everything so we can efficiently get through all
the issues and PRs. But rest assured, we are working hard and even working on some partnerships behind the scenes to really help this project take off!
### How do local LLMs fair compared to larger models like Claude 3.5 Sonnet for oTToDev/Bolt.new?
As much as the gap is quickly closing between open source and massive close source models, you’re still going to get the best results with the very large models like GPT-4o, Claude 3.5 Sonnet, and DeepSeek Coder V2 236b. This is one of the big tasks we have at hand - figuring out how to prompt better, use agents, and improve the platform as a whole to make it work better for even the smaller local LLMs!
### I'm getting the error: "There was an error processing this request"
If you see this error within oTToDev, that is just the application telling you there is a problem at a high level, and this could mean a number of different things. To find the actual error, please check BOTH the terminal where you started the application (with Docker or pnpm) and the developer console in the browser. For most browsers, you can access the developer console by pressing F12 or right clicking anywhere in the browser and selecting “Inspect”. Then go to the “console” tab in the top right.
### I'm getting the error: "x-api-key header missing"
We have seen this error a couple times and for some reason just restarting the Docker container has fixed it. This seems to be Ollama specific. Another thing to try is try to run oTToDev with Docker or pnpm, whichever you didn’t run first. We are still on the hunt for why this happens once and a while!
### I'm getting a blank preview when oTToDev runs my app!
We promise you that we are constantly testing new PRs coming into oTToDev and the preview is core functionality, so the application is not broken! When you get a blank preview or don’t get a preview, this is generally because the LLM hallucinated bad code or incorrect commands. We are working on making this more transparent so it is obvious. Sometimes the error will appear in developer console too so check that as well.
This goes to the point above about how local LLMs are getting very powerful but you still are going to see better (sometimes much better) results with the largest LLMs like GPT-4o, Claude 3.5 Sonnet, and DeepSeek Coder V2 236b. If you are using smaller LLMs like Qwen-2.5-Coder, consider it more experimental and educational at this point. It can build smaller applications really well, which is super impressive for a local LLM, but for larger scale applications you want to use the larger LLMs still!
