I'm building something BIG behind the scenes - a much larger platform that is also going to have our Discourse community for oTToDev within it but also a LOT more for all of you! ottomator.ai/
Can we somehow pull cofounder into the build as well or fork the project. I really like how it has the wireframes but I wanna use it with Ollama. Idk if that would be hard or not
@@matthewcallaham1024 Cofounder is very different from Bolt. Cofounder plans everything in advance while Bolt is more focused on iterating over a vague initial idea
Great work Colin, I do hope that there will be some info about what models that can be run locally, accelerated by gpu, that works well with ottodev. I'd rather have to ask more prompts to get the right results that to pay a lot for every prompt to providers.
Great job! Thank you Cole and all the contributors for making this happen. Keeping pushing this project to its limit. I haven't contributed anything so far to this project, however I'm working on something and once I could able to monetize it, I'll make sure that I'll hire a resource exclusively to work on this project and maintain it.
Thanks for the redirection to this video for the full install, it wasnt found by me easily either. Perhaps a stand alone video for the installation instructions would be gr8 for engaging more newbies
Keep going Man! We are watching you from the beginning...I think you are gonna break bolt.new users ..haha..just waiting for twist..for my work pressure, I can't join in the contributor, But I really going to support you man...bOOOOOMMMMMMMm
I found about your YT Chanel through the Git repo. Awesome work, keep it up. Btw, maybe this project can be expanded to use AI agents, where you could create a plan/specifications with an AI agent, and then have a back-end and a front-end and designer AI Agents collaborating to make a full stack app.
Thank you, I sure will keep it up! I LOVE it - agents are certainly something I want to implement as an extension to this, maybe even as an optional feature to try to get better results at the cost of some more tokens.
ok these problems should have solution : 1-left sidebar should disable it for automatic toggle 2-add gpu for docker 3-make it more free to all poor users for me example i tried 20 different os , installation guide so clean and perfect , thanks
New Sub, falling down the rabbit hole. I'm highly interested in joining in your bolt project. I have the time but am trying to get up to speed. Thanks for the great content
Awesome job on the project and the video, I have it up and running. Thanks so much Cole, I asked below but I'm going to ask here also could you show us how to update it using docker. Also I wish I could help out but I'm just learning to code.
You are so welcome, I'm glad you have it up and running! To update with Docker, all you have to do is pull the latest changes from the GitHub repository and then restart your containers! "docker compose down" will stop the containers and you can rebuild/rerun to make sure you have the latest changes with the same commands you used to start it initially.
thank you for thanks for your commitment to driving this cool project forward. are there plans to integrate further programming languages or programming environments with different frameworks (different frontend and backend)?
Hey thanks for the upload just saw one of you videos yesterday and was confused with installation and today you uploaded the instructions hahaha , I will be doing it later in my free time , Just had one question can we integrate Claude with Xcode to perform something similar to cursorai ?
Perfect timing, let me know how it goes! Yes you certainly could! Essentially making an XCode extension that uses Claude to edit your code or suggest edits!
I'm so pumped up for this project and I want to contribute but don't know typescript. But I do have DevOps experience. Is there anyway I can contribute to the repository?
Thank you and if there is anything you can think of at all for contributing just let me know! There are always ways to help even if you don't know Typescript, I'm open to whatever you think of!
So, I had the idea to work on something exactly like Bolt, and I found bolt today, and today I also found your fork, isn't that too crazy of a coincidence?
Linux has Chrome binaries available. Also the open source version: Chromium. I use Brave which has the chromium code base in it. The only big fork is that chrome removed some ability to perform adblocking. Brave didn't use this feature so you can block ads with Brave.
Is it possible to add a feature like cofounder ai? I think it would be the better open source coding tool if it can get to work like cofounders using wireframes and a visual way to see how everything is connected
Hmm, if you went for docker was it not better to install node inside of it and not on your system? I know people struggled with node install in windows. We probably should make docker image so that whole process is install docker and download and run image.
Yeah node is installed in the image! I just recommend installing Node on the host machine as well since there are some nice NPM scripts the pull request author for this added in to run some of the Docker commands easier.
Hey Cole, there are problems with Ollama, whenever i try sending a message, i get a notification saying there was an error sending your message. Plus, i can’t see anything on preview for some reason. A simple app video showing us how you make the repo work , how to install would be really helpful. Remember there are so many people coming from no code perspective
I made some fixes for Ollama over the past couple of days that should help! Also some of the smaller models don't handle the Bolt.new prompting the best so I'd use a larger one if you are using 3b parameters or less. More content coming for tutorials soon too!
Is this new fork fixing any of the issues like database permission errors, image upload permission denied, and most importantly, modifying a project only for it to lose context and completely rewrite the file or forget to add functions or features back that previously existed? These are the main issues we need fixed to make this usable.
@ColeMedin Say I'm building an app for diet tracking. Then I decide I want to be able to add new foods to a database that will be searchable like myfitnesspal. I'm able to add foods to my diary, delete foods, edit calories etc but when I want to add a new feature like a food database, firebase should create a new collection called say food Database. It constantly gets permission denied when trying to create new fields or collections. I've prob fixed 700 errors and 10M tokens for it still to not work. Im a pretty heavy user of bolt. Prob 15M tokens a day right now.
a warning for those who know nothing about programming: fork of bolt does not want to work with free models, it causes hundreds of errors when working with free models and it will not fix its own errors. it does not give you the preview of your app and you will not find help in the community of this fork (knowledgeable people ignore everyone, and noobs can't help each other) If you have at least a basic knowledge of programming or are willing to pay for paid models, this fork is perfect for you.
Free and smaller LLMs definitely can struggle with Bolt.new in general, so I am with you there! The Bolt.new prompt behind the scenes is pretty complicated and requires the LLM to output things in a specific way. But the community is growing and once we have a Discourse there will always be a place for people to get help if they want!
You just saved me some serious time 😅... Is it worth getting oTTo to create the file structure then take it over to ChatGPT or whatever for fixing? Even the online official bolt ,new. Or is passing code arround models not advisable.
@@ColeMedin thank you for your work I really hope that the community will start helping newbies I'm using the main bolt for now, but I'm following your fork daily hoping to start using it soon :)
Cole Thanks very much for this. Have two questions regarding the setup if you can please assist. Question 1 - regarding the installation with Docker. in Step 1 it brings up warnings related to the API's "SecretsUsedInArgOrEnv: Do not use ARG or ENV instructions for sensitive data. In your video it was also the case and I see in your notes you mentioned not to commit the .Env file to version control and its already in the .gitignore. is that sufficient or should we use one of the other methods for storing these? Docker Secrets /environment variables outside docker file / build kit to hide secrets / Docker compose secrets? Question 2 - After the development stages of the projects. Do you then need to run the Docker build production & Docker compose production. Then for each project toggle between the two states?
You are welcome! 1. Not committing the .env is sufficient! I just also wouldn't push the image to DockerHub or something after you build it, just keep it local. 2. You just need to run the production build commands in the environment you want to run oTToDev in production!
Hello ! I encounter a problem. I followed all the steps and the tool is showing on my browser in local but when I try to ask something I get this message : "There was an error processing your request". Does anyone have an idea why and how to solve it ?
@@ColeMedin And the part 3 : bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' Bolt creates a SINGLE, comprehensive artifact for each project. The artifact contains all necessary steps and components, including: ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' - Shell commands to run including dependencies to install using a package manager (NPM) ' + bolt-ai-dev-1 | ' - Files to create and their contents ' + bolt-ai-dev-1 | ' - Folders to create if necessary ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' 1. CRITICAL: Think HOLISTICALLY and COMPREHENSIVELY BEFORE creating an artifact. This means: ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' - Consider ALL relevant files in the project ' + bolt-ai-dev-1 | ' - Review ALL previous file changes and user modifications (as shown in diffs, see diff_spec) ' + bolt-ai-dev-1 | ' - Analyze the entire project context and dependencies ' + bolt-ai-dev-1 | ' - Anticipate potential impacts on other parts of the system ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' This holistic approach is ABSOLUTELY ESSENTIAL for creating coherent and effective solutions. ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' 2. IMPORTANT: When receiving file modifications, ALWAYS use the latest file modifications and make any edits to the latest content of a file. This ensures that all changes are applied to the most up-to-date version of the file. ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' 3. The current working directory is `/home/project`. ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' 4. Wrap the content in opening and closing `` tags. These tags contain more specific `` elements. ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' 5. Add a title for the artifact to the `title` attribute of the opening ``. ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' 6. Add a unique identifier to the `id` attribute of the of the opening ``. For updates, reuse the prior identifier. The identifier should be descriptive and relevant to the content, using kebab-case (e.g., "example-code-snippet"). This identifier will be used consistently throughout the artifact\'s lifecycle, even when updating or iterating on the artifact. ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' 7. Use `` tags to define specific actions to perform. ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' 8. For each ``, add a type to the `type` attribute of the opening `` tag to specify the type of the action. Assign one of the following values to the `type` attribute: ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' - shell: For running shell commands. ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' - When Using `npx`, ALWAYS provide the `--yes` flag. ' + bolt-ai-dev-1 | ' - When running multiple shell commands, use `&&` to run them sequentially. ' + bolt-ai-dev-1 | ' - ULTRA IMPORTANT: Do NOT re-run a dev command if there is one that starts a dev server and new dependencies were installed or files updated! If a dev server has started already, assume that installing dependencies will be executed in a different process and will be picked up by the dev server. ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' - file: For writing new files or updating existing files. For each file add a `filePath` attribute to the opening `` tag to specify the file path. The content of the file artifact is the file contents. All file paths MUST BE relative to the current working directory. ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | " 9. The order of the actions is VERY IMPORTANT. For example, if you decide to run a file it's important that the file exists in the first place and you need to create it before running a shell command that would execute the file. " + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' 10. ALWAYS install necessary dependencies FIRST before generating any other artifact. If that requires a `package.json` then you should create that first! ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' IMPORTANT: Add all required dependencies to the `package.json` already and try to avoid `npm i ` if possible! ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' 11. CRITICAL: Always provide the FULL, updated content of the artifact. This means: ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' - Include ALL code, even if parts are unchanged ' + bolt-ai-dev-1 | ' - NEVER use placeholders like "// rest of the code remains the same..." or "" ' + bolt-ai-dev-1 | ' - ALWAYS show the complete, up-to-date file contents when updating files ' + bolt-ai-dev-1 | ' - Avoid any form of truncation or summarization ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' 12. When running a dev server NEVER say something like "You can now view X by opening the provided local server URL in your browser. The preview will be opened automatically or by the user manually! ' + bolt-ai-dev-1 | ' ' + bolt-ai-dev-1 | ' 13. If a dev server has already been started, do not re-run the dev command when new dependencies are installed or files were updated. Assu'... 5141 more characters, bolt-ai-dev-1 | messages: [ [Object] ], bolt-ai-dev-1 | tools: undefined, bolt-ai-dev-1 | tool_choice: undefined, bolt-ai-dev-1 | stream: true bolt-ai-dev-1 | }, bolt-ai-dev-1 | statusCode: 401, bolt-ai-dev-1 | responseHeaders: { bolt-ai-dev-1 | 'cf-cache-status': 'DYNAMIC', bolt-ai-dev-1 | 'cf-ray': '8dd785a5b844e209-MRS', bolt-ai-dev-1 | connection: 'keep-alive', bolt-ai-dev-1 | 'content-length': '97', bolt-ai-dev-1 | 'content-type': 'application/json', bolt-ai-dev-1 | date: 'Mon, 04 Nov 2024 20:58:25 GMT', bolt-ai-dev-1 | 'request-id': 'req_01J29jc9wfmTRXHxRMvZvvto', bolt-ai-dev-1 | server: 'cloudflare', bolt-ai-dev-1 | via: '1.1 google', bolt-ai-dev-1 | 'x-robots-tag': 'none', bolt-ai-dev-1 | 'x-should-retry': 'false' bolt-ai-dev-1 | }, bolt-ai-dev-1 | responseBody: '{"type":"error","error":{"type":"authentication_error","message":"x-api-key header is required"}}', bolt-ai-dev-1 | isRetryable: false, bolt-ai-dev-1 | data: { bolt-ai-dev-1 | type: 'error', bolt-ai-dev-1 | error: { bolt-ai-dev-1 | type: 'authentication_error', bolt-ai-dev-1 | message: 'x-api-key header is required' bolt-ai-dev-1 | } bolt-ai-dev-1 | }, bolt-ai-dev-1 | [Symbol(vercel.ai.error)]: true, bolt-ai-dev-1 | [Symbol(vercel.ai.error.AI_APICallError)]: true bolt-ai-dev-1 | }
@@ColeMedin BTW, thank you for your response ! After several hours of struggle, I give up for today. I'll check your answer tomorrow and hope there will be a solution. PS : Thanks for your work, I don't know for me yet but I saw that a lot of people are very happy using this product.
Awesome, I configured it on local machine with Docker but seems its not working when selecting ollama with qwen model, it seems to be taking Anthropic Model by default so ignoring Ollama.
@@ColeMedin I think that if you don't set up an API for Anthropic the tool won't work. I have the same problem (no anthropic pro account or credits for the API) and when I try to ask something I got the same error as everyone else. When I check the logs I got a message saying something about not being able to connect to Anthropic API. I also put an API but as I already mentioned, without credits, and the log message is different, saying something about "not enough credits on Anthropic". All these while choosing Ollama on the UI menu that you created for the tool.
Thanks man this stuff is awesome wish you luck, at some point I might add some of my web scraping with llm scripts into this , I dont do node js mostly python but you know I am sure their is a way this is gonna be the best 😃Pls upvote
Good question! All you have to do is run the git pull (either in the command line or in the Git extension for VSCode) and then restart the site/container to make sure the latest changes are reflected!
I have a question: I can only get this to jive with Ollama if I access it via localhost:5173. If I attempt to access it via the IP address (ie: 10.0.0.166:5173) I can access the frontend, but it will no longer display, or use, the Ollama models. This seems to be whether Ollama is hosted on the same machine as ottodev, or on another instance of ollama on the network (in this case, my ollama server). Both access methods require that I use the frontend at localhost, strictly, but I would prefer to run the frontend on its own server, if possible. Any suggestions?
Thanks for the walkthrough on setting Ottodev up with Docker. Got it all up and running but it can't see any of my API keys. For each model where I've set the API key in .env I see the following error as per the OpenAI example here... "OpenAI API Key: Not set". Would appreciate some assistance - is the Discource community up and running yet?
Awesome video! I’m really excited to take on a challenge where I’ll develop an AI startup idea and launch it in just 48 hours starting on November 7. I want to see what I can create in such a short time! I’m still on the lookout for the perfect idea, so I’d love to hear any suggestions from you all. I’m aiming for something fun, a bit challenging, but still manageable within 48 hours. Please share any ideas you have-I’ll choose one and dive right in. Thanks in advance for your input, and let’s see what unfolds!
Once I complete the AI project, I’ll share a video outlining the entire process I went through to create it. So buckle up-this journey is going to be exciting!
Everything works good but except one thing, I entered all api key for all llm but since they are in free version i cant use them in docker, also, when sending a message to the only one working, it always repeat what the first message was saying, without seeing next messages
As much as I would love to try this I am getting nothing but problems following the directions. I try the "npm run dockerbuild:prod" but get "ERROR: "docker buildx build" requires exactly 1 argument." So then I try the "docker build . --target bolt-ai-production" and then "docker-compose --profile production up" But I always end up with ".env.local: no such file or directory" Im wondering at what point does the .env.local get created.
even him dosent know what hes talking about, he says in the video to create .env file but then the bindings script for docker build is looking for .env.local .... : cat bindings.sh #!/bin/bash bindings="" while IFS= read -r line || [ -n "$line" ]; do if [[ ! "$line" =~ ^# ]] && [[ -n "$line" ]]; then name=$(echo "$line" | cut -d '=' -f 1) value=$(echo "$line" | cut -d '=' -f 2-) value=$(echo $value | sed 's/^"\(.*\)"$/\1/') bindings+="--binding ${name}=${value} " fi done < .env.local bindings=$(echo $bindings | sed 's/[[:space:]]*$//') echo $bindings .... in my case i cant even manage to make it work with the commands in the git readme file .... pitty ... save your time and run it locally : pnpm run build && pnpm run start
Great question! I believe you asked this in another video, but around ~$0.02 on average I would say. Given Claude 3.5 Sonnet is $3 per million input and $15 per million output tokens.
hey, a non dev here. Followed the docker set up but environment 'usually' does not build a preview in the ui giving a "You can now preview the updated homepage by running the development server." How do I rectify? Using gpt4omini (as have a bunch of tokens) - seems happy to chat but doesn't add much code.
That's weird! So sometimes everything works great but other times it doesn't? I'm guessing that is a limitation of GPT-4o-mini not executing the right commands versus it actually being an issue within the application. Unless you try with something like Claude 3.5 Sonnet and it still happens often!
I've got a big ask. Where do we start?? There so many places trying to say how to... Do you have any tips? I'm also looking at a new AI laptop? I'm all in!! Thanks a bunch
Great question - I love your enthusiasm! Are you asking where to start as far as contributing? That's really up to you and how you want to help improve the project! And I'll also be putting out a video soon on how to contribute and what that looks like!
Sometimes it can take a while for the preview to show up... but if it isn't showing up even after a while, you can open the terminal with the button in the top right and then try to run the command yourself. It's weird that sometimes the LLM will also not give the right command so that could be possible too.
Hi Cole! I just followed this video and installed everything and I'm using Gemini 1.5 Pro api. However, no matter what I enter in the prompt, I get a white screen in Preview. I'm also using Canary as you mentioned with no extensions installed. How can I troubleshoot this?
So that usually means there is an error with the LLM hallucinating a bad command or bad code. I cover troubleshooting steps in this video (timestamp included in link): th-cam.com/video/5I9VgwauuzU/w-d-xo.html
This video says to use .env, they docker on Linux won't run unless the file is called .env.local. Even then it doesn't seem to read the Enviornment. OpanAI won't work.I am excited to be able to use this.
From what I understand Docker containers always use .env by default no matter the OS... did you test this yourself to be sure? I'm super curious, might have to update the docs!
Thanks Cole for the great work, I don't know where i have gone wrong but I keep getting this error, "There was an error processing your request" on the chat.
You are welcome! That error happens once and a while... what is the error message you see in the terminal where you started the site? I'd also check the developers tools (F12) in the browser and see the errors in the console there too.
hey man no matter what i do i just seem to keep on getting the same results and that is the " There was an error processing your request; no details were returned" . Please let me know what i am doing wrong as there is nothing more frustrating than trying multiple different routes and still just being caught up at the same error. thanks man i really do think what your doing is awesome by the way and i hope this project really takes off like it should.
Thank you! Sorry you are running into this error! What is the error message you see in the terminal where you started the site, or in the developer console in the browser? Typically that is much more helpful.
Great question! It's a play on words to mean automated developer kind of as you said! The name also ties into something bigger behind the scenes I'm building for all of us!
Those Devin like apps try to do a lot for you, usually way too much haha oTToDev and Bolt.new is specifically meant to generate full stack web apps in the browser and it does that super well.
I follow exactly as you showed in this video. But I got this error message: "There was an error processing your request" There was API key setup problem. I was using Ollama that runs on my computer. I am not sure how to fix the error.
Sorry you're running into this! What is the error message you get in the terminal where you started the container? Usually that error message helps a lot more than the one on the frontend you referenced.
@@ColeMedin Buried in the text resulting from "npm run dockerbuild" is "| APICallError [AI_APICallError]: x-api-key header is required". I've renamed the file .env with only this change to its contents: # You only need this environment variable set if you want to use oLLAMA models # EXAMPLE localhost:11434 OLLAMA_API_BASE_URL=localhost:11434 I'm on a Mac, if I go to localhost:11434 in Google Chrome Canary I get the message "Ollama is running" and the downloaded LLM I have is llama3.2 Can you see where the error is?
@ColeMedin Here is the error: "APICallError [AI_APICallError]: x-api-key header is required" **When I used OpenAI API key it works.** **But having error when using Ollama API URL.** My PC Setup: - My PC OS is Windows 11. - Installed Ubuntu OS for Windows. - Then, installed Docker in Ubuntu and then Ollama with Openweb UI, and multiple llm models. - It's working with no problem. *Then I saw your video(This one) and followed your instructions* - Clone the repo - Installed Docker Desktop on Windows and so on. - It's working with OpenAI API Key. -* But having problem with Ollama.*
I'm having trouble getting the preview to work the run npm just stalls out cause the preview shares the same host port .. not sure I'm trying to claude troubleshoot this but im total noob I was able to get docker and container set up fine bolt working great but no preview.. tried to fix it for hours last night Might try the pinnocio method or just get the stax service for containers
Even if I run it with docker I can't get it working for preview either.. I think the issue is that I might have had a project open on my phone that was using the same port so im gonna try to work around that now and see if it fixes
@ColeMedin - i'm pretty new to this, but have managed to setup the oTTo/Bolt environment based on your instructions. Thanks. I'm also using it with Ollama. So cool. However, when I ask this local Bolt to create a someething (such as a basic 'hello world' html page) it prints out the code in the chat, but doesn't create the project files, run it in preview etc. on the right side of the screen. Any idea why this might not be running the way the original Bolt works. I must have done something wrong in my setup. Thanks
Actually, I've seen your latest video and this has answered my questions. The only problem now, is that my external API-Keys don't appear to be working. Not sure why as I have updated constants.ts and .env
@@ColeMedin it's something to do with my api setup. none of the api's i've setup (openrouter, openai, deepseek etc.) will work. my local Ollama works, just not the LLMs that I connect to via api key. whenever i try to communicate to the LLMs via the chatbox, a message pops up on the right of the screen "There was an error processing your request"
I got it everything running but when I type make something I get an error there was an error pressing your request and it goes the same for all the api I add it looks like somehow the app can’t call the api
Fair question! I'm not totally sure to be honest, even with the base open source Bolt.new you'll just get an error with Chrome saying it doesn't work with that browser and to use Canary. I believe it's something with the WebContainer that doesn't work with Chrome when you're running the app locally.
Thanks for the reply yeah what's your working on is pretty awesome and groundbreaking The reason why I ask is because I like the mobility of the smartphone being able to work on it everywhere and I understand the power limitations compared to a laptop or a PC but of course these things are limitations at first.
I still need help with installing this. I have watched the video multiple times and read the notes back to back. I wonder if the problem is with Apple Silicon or me. Can anyone help me install this? I would be very grateful.
Also, please don't switch it up and start charging once you get to the juicey bits. I know we all have to make a living. But nothing worse than looking for a bootstrap, getting three quaters through a video and find out you have to pay. If its like a one off $49.99 or something, maybe. But I'm not looking to sign up per month. I admire your work, trust me. But please try and keep it for the poor in the cost of living crisis. Thats why we are here. 😉
My plan is to always keep this open source and free to use! I never want to start adding commercial features that you need to pay for like Bolt.new is doing, though I understand why they do that. That isn't for oTToDev though!
@@ColeMedin I was more referrring to when I signed up to your website, I got anxiety lol. No Im kinding, your'e good at what you do, just please, try and keep it low.
I talk about it more in the video but it'll be good to give it a distinct name now that the project is evolving heavily instead of just calling it "the fork". Plus the name plays really well into the larger platform that I am working on behind the scenes for everyone!
Hi Cole, I am running into a ton of dependency errors when trying to run both on docker and on my OS locally. Despite over five attempts, I can't seem to get my config correct with the API keys. It either pulls the local instances or fails to map them correctly to the dropdowns. I can share what I have so far and debug further, I just don't have visibility if anyone else has this issue. Please let me know how to best connect either 1:1 or in the discord, ideally before this coming sunday.
Sorry you're running into these issues! Is this for Ollama specifically? I actually made a fix yesterday for this that might help, so I'd try repulling the repo and running again!
I'm building something BIG behind the scenes - a much larger platform that is also going to have our Discourse community for oTToDev within it but also a LOT more for all of you!
ottomator.ai/
Can we somehow pull cofounder into the build as well or fork the project. I really like how it has the wireframes but I wanna use it with Ollama. Idk if that would be hard or not
We need background IT agents team for better coordination and prompt caching to reduce api costs..
Already Signed up for this, Really Excited about this Project: )
@@ColeMedin when is launch day
@@matthewcallaham1024 Cofounder is very different from Bolt. Cofounder plans everything in advance while Bolt is more focused on iterating over a vague initial idea
Congrats man. Other channels are covering your fork, wish you all the best for providing value to the community
Thanks man! Yeah it's exciting!
Dude your genuine excitement and gratitude is awesome to see! Keep going bro!!
Thanks man! I appreciate it a ton!
You are going to famous soon bro.....just grind this month and luck is on your way for sure.
I appreciate it man!
Great work Colin, I do hope that there will be some info about what models that can be run locally, accelerated by gpu, that works well with ottodev. I'd rather have to ask more prompts to get the right results that to pay a lot for every prompt to providers.
a lot of youtubers i follow in this Ai-journey are giving shoutout to you. Its amazing to be a part of witnessing the growth. Keep it up bro!
Yeah it's fantastic! Thanks man!!
Great job! Thank you Cole and all the contributors for making this happen. Keeping pushing this project to its limit. I haven't contributed anything so far to this project, however I'm working on something and once I could able to monetize it, I'll make sure that I'll hire a resource exclusively to work on this project and maintain it.
Sounds fantastic, thank you!!
Great video bro.I look forward to new videos from you every week.Thanks for sharing your knowledge with us..
You are so welcome - thank you!
Thanks for the redirection to this video for the full install, it wasnt found by me easily either. Perhaps a stand alone video for the installation instructions would be gr8 for engaging more newbies
You bet! Yeah that's fair!
Keep going Man! We are watching you from the beginning...I think you are gonna break bolt.new users ..haha..just waiting for twist..for my work pressure, I can't join in the contributor, But I really going to support you man...bOOOOOMMMMMMMm
Docker is really the key to making this accessible to all I believe. this makes it sooo easy
You are Awesome... cant wait to see your Video on this updates. Kindly make it a step by step for us Newbies to Properly Implement it, LFG : )
I found about your YT Chanel through the Git repo. Awesome work, keep it up. Btw, maybe this project can be expanded to use AI agents, where you could create a plan/specifications with an AI agent, and then have a back-end and a front-end and designer AI Agents collaborating to make a full stack app.
Thank you, I sure will keep it up! I LOVE it - agents are certainly something I want to implement as an extension to this, maybe even as an optional feature to try to get better results at the cost of some more tokens.
ok these problems should have solution :
1-left sidebar should disable it for automatic toggle
2-add gpu for docker
3-make it more free to all poor users
for me example i tried 20 different os , installation guide so clean and perfect , thanks
Thanks for the suggestions - and please feel free to contribute if you have solutions!
Glad the installation is working well for you!
New Sub, falling down the rabbit hole. I'm highly interested in joining in your bolt project. I have the time but am trying to get up to speed. Thanks for the great content
Awesome job on the project and the video, I have it up and running. Thanks so much Cole, I asked below but I'm going to ask here also could you show us how to update it using docker. Also I wish I could help out but I'm just learning to code.
You are so welcome, I'm glad you have it up and running!
To update with Docker, all you have to do is pull the latest changes from the GitHub repository and then restart your containers! "docker compose down" will stop the containers and you can rebuild/rerun to make sure you have the latest changes with the same commands you used to start it initially.
thank you for thanks for your commitment to driving this cool project forward.
are there plans to integrate further programming languages or programming environments with different frameworks (different frontend and backend)?
You are so welcome! That is on the long term roadmap!
Well done mate this is phenomenal work. Cant wai to start building
Great to see this project evolving massively!😊
Thank you for this amazing video. This video should have millions like. Thank you. You're doing an amazing job with this project.
Thank you so much, I appreciate it a ton!! :D
Hey thanks for the upload just saw one of you videos yesterday and was confused with installation and today you uploaded the instructions hahaha , I will be doing it later in my free time , Just had one question can we integrate Claude with Xcode to perform something similar to cursorai ?
Perfect timing, let me know how it goes!
Yes you certainly could! Essentially making an XCode extension that uses Claude to edit your code or suggest edits!
I'm so pumped up for this project and I want to contribute but don't know typescript. But I do have DevOps experience. Is there anyway I can contribute to the repository?
Thank you and if there is anything you can think of at all for contributing just let me know! There are always ways to help even if you don't know Typescript, I'm open to whatever you think of!
Really good to train this so it used a simple Database such as PocketBase as alternative to localstore. This would fast track to MVP.
So, I had the idea to work on something exactly like Bolt, and I found bolt today, and today I also found your fork, isn't that too crazy of a coincidence?
That is crazy!
thank you sir! i have no clue how to do this. but im reading and watching! :D
Great intro for newbies. I'll be sharing this video to get people started. Interesting that it doesn't work in Chrome.
Thank you!! Yeah it is weird, Chrome Canary is easy to install though luckily.
Great video for us noobs, thank you! 🙏
Looking forward to Sunday. 🚀
You are so welcome!
Linux has Chrome binaries available. Also the open source version: Chromium. I use Brave which has the chromium code base in it. The only big fork is that chrome removed some ability to perform adblocking. Brave didn't use this feature so you can block ads with Brave.
By the way, could you schedule the live stream for Sunday on TH-cam so we can enable the notification for it?
Thank you for suggesting that! I have it scheduled now:
th-cam.com/video/_YzTntvUWN4/w-d-xo.html
I wish i was good at coding to be able help you guys build! next feature needs attachments and uploads asap lol
I agree!
Loved it, But one feature is missing as Volt is the uploading picture. I can't upload picture in the oTTodev
Thanks and yeah that is in development! It's not in the open source version of Bolt.new so we have to implement it ourselves!
Is it possible to add a feature like cofounder ai? I think it would be the better open source coding tool if it can get to work like cofounders using wireframes and a visual way to see how everything is connected
Fantastic question! Something that quite a few individuals have brought up, so I certainly have this on my list to look into!
Hmm, if you went for docker was it not better to install node inside of it and not on your system?
I know people struggled with node install in windows.
We probably should make docker image so that whole process is install docker and download and run image.
Yeah node is installed in the image! I just recommend installing Node on the host machine as well since there are some nice NPM scripts the pull request author for this added in to run some of the Docker commands easier.
Let's integrate both voice assistant creation tool like (synth flow or vapi) and chatbot creation tool like voice
flow into it.
That would be sweet!
Hey Cole, there are problems with Ollama, whenever i try sending a message, i get a notification saying there was an error sending your message. Plus, i can’t see anything on preview for some reason. A simple app video showing us how you make the repo work , how to install would be really helpful. Remember there are so many people coming from no code perspective
I made some fixes for Ollama over the past couple of days that should help! Also some of the smaller models don't handle the Bolt.new prompting the best so I'd use a larger one if you are using 3b parameters or less.
More content coming for tutorials soon too!
bolt vs replit... or could we combine the best parts of each?😅
Definitely possible to combine them, I love that though!
Is this new fork fixing any of the issues like database permission errors, image upload permission denied, and most importantly, modifying a project only for it to lose context and completely rewrite the file or forget to add functions or features back that previously existed? These are the main issues we need fixed to make this usable.
These are all things being worked on! Except could you clarify the database permission errors? I'm not familiar with that one.
@ColeMedin Say I'm building an app for diet tracking. Then I decide I want to be able to add new foods to a database that will be searchable like myfitnesspal. I'm able to add foods to my diary, delete foods, edit calories etc but when I want to add a new feature like a food database, firebase should create a new collection called say food Database. It constantly gets permission denied when trying to create new fields or collections. I've prob fixed 700 errors and 10M tokens for it still to not work.
Im a pretty heavy user of bolt. Prob 15M tokens a day right now.
@@ColeMedin im wondering if we could add something like sqllight or pocketbase for full stack. It would allow bolt to easily read the files
sounds familiar. it's sounding like firestore permission issue. Check your settings in firebase did not help?
Yeah I love the suggestion @ContentVibeio!
Love your work. Is it possible to import projects into ottodev
Thanks! There is a PR open for this that I'm going to take a look for soon actually!
@ColeMedin Thats awesome. I have just started to learn to code. Hopefully I can contribute something 😁
Sounds fantastic, I would love that!
Great work! Thanks for the clear, step-by-step guidance!
You're welcome!
I was hoping to get a mention for helping you on install docs 😂. Great work on this project..
Shoot you're totally right I should have! My bad! I will be sure to update the README for that
Are you @hillct on GitHub? If not he helped with the docs a lot as well. What is your GitHub username?
Hi Can you get the merge request closed? Changes are long pending. If you need any help with that please let us know.
I appreciate it! I'm working on getting more people involved to help with all the pull requests as we speak!
a warning for those who know nothing about programming:
fork of bolt does not want to work with free models, it causes hundreds of errors when working with free models and it will not fix its own errors.
it does not give you the preview of your app and you will not find help in the community of this fork (knowledgeable people ignore everyone, and noobs can't help each other)
If you have at least a basic knowledge of programming or are willing to pay for paid models, this fork is perfect for you.
Free and smaller LLMs definitely can struggle with Bolt.new in general, so I am with you there! The Bolt.new prompt behind the scenes is pretty complicated and requires the LLM to output things in a specific way.
But the community is growing and once we have a Discourse there will always be a place for people to get help if they want!
You just saved me some serious time 😅... Is it worth getting oTTo to create the file structure then take it over to ChatGPT or whatever for fixing? Even the online official bolt ,new. Or is passing code arround models not advisable.
@@ColeMedin thank you for your work
I really hope that the community will start helping newbies
I'm using the main bolt for now, but I'm following your fork daily hoping to start using it soon :)
Cole Thanks very much for this. Have two questions regarding the setup if you can please assist.
Question 1 - regarding the installation with Docker. in Step 1 it brings up warnings related to the API's "SecretsUsedInArgOrEnv: Do not use ARG or ENV instructions for sensitive data. In your video it was also the case and I see in your notes you mentioned not to commit the .Env file to version control and its already in the .gitignore. is that sufficient or should we use one of the other methods for storing these? Docker Secrets /environment variables outside docker file / build kit to hide secrets / Docker compose secrets?
Question 2 - After the development stages of the projects. Do you then need to run the Docker build production & Docker compose production. Then for each project toggle between the two states?
You are welcome!
1. Not committing the .env is sufficient! I just also wouldn't push the image to DockerHub or something after you build it, just keep it local.
2. You just need to run the production build commands in the environment you want to run oTToDev in production!
We made it! I just saw another AI channel, WorldofAI cover ottoDev
I love it!! Thanks for mentioning that!
Thanks cole for update ❤️
Bolt is adding all you implemented on theirs
That's how opensource works, we are using their base let's not forget that.
Yeah I'm totally good with that! As @Jonathan-o1h9x that's how open source works so I'm game!
Hello ! I encounter a problem. I followed all the steps and the tool is showing on my browser in local but when I try to ask something I get this message :
"There was an error processing your request". Does anyone have an idea why and how to solve it ?
i have same problem. Please Help :) thanks
Sorry you're running into that! What is the error you are seeing in the terminal where you ran the Docker container?
@@ColeMedin And the follow up of the error : bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' Available shell commands:
' +
bolt-ai-dev-1 | ' File Operations:
' +
bolt-ai-dev-1 | ' - cat: Display file contents
' +
bolt-ai-dev-1 | ' - cp: Copy files/directories
' +
bolt-ai-dev-1 | ' - ls: List directory contents
' +
bolt-ai-dev-1 | ' - mkdir: Create directory
' +
bolt-ai-dev-1 | ' - mv: Move/rename files
' +
bolt-ai-dev-1 | ' - rm: Remove files
' +
bolt-ai-dev-1 | ' - rmdir: Remove empty directories
' +
bolt-ai-dev-1 | ' - touch: Create empty file/update timestamp
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' System Information:
' +
bolt-ai-dev-1 | ' - hostname: Show system name
' +
bolt-ai-dev-1 | ' - ps: Display running processes
' +
bolt-ai-dev-1 | ' - pwd: Print working directory
' +
bolt-ai-dev-1 | ' - uptime: Show system uptime
' +
bolt-ai-dev-1 | ' - env: Environment variables
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' Development Tools:
' +
bolt-ai-dev-1 | ' - node: Execute Node.js code
' +
bolt-ai-dev-1 | ' - python3: Run Python scripts
' +
bolt-ai-dev-1 | ' - code: VSCode operations
' +
bolt-ai-dev-1 | ' - jq: Process JSON
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' Other Utilities:
' +
bolt-ai-dev-1 | ' - curl, head, sort, tail, clear, which, export, chmod, scho, hostname, kill, ln, xxd, alias, false, getconf, true, loadenv, wasm, xdg-open, command, exit, source
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' Use 2 spaces for code indentation
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' You can make the output pretty by using only the following available HTML elements: , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' For user-made file modifications, a `` section will appear at the start of the user message. It will contain either `` or `` elements for each modified file:
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' - ``: Contains GNU unified diff format changes
' +
bolt-ai-dev-1 | ' - ``: Contains the full new content of the file
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' The system chooses `` if the diff exceeds the new content size, otherwise ``.
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' GNU unified diff format structure:
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' - For diffs the header with original and modified file names is omitted!
' +
bolt-ai-dev-1 | ' - Changed sections start with @@ -X,Y +A,B @@ where:
' +
bolt-ai-dev-1 | ' - X: Original file starting line
' +
bolt-ai-dev-1 | ' - Y: Original file line count
' +
bolt-ai-dev-1 | ' - A: Modified file starting line
' +
bolt-ai-dev-1 | ' - B: Modified file line count
' +
bolt-ai-dev-1 | ' - (-) lines: Removed from original
' +
bolt-ai-dev-1 | ' - (+) lines: Added in modified version
' +
bolt-ai-dev-1 | ' - Unmarked lines: Unchanged context
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' Example:
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' @@ -2,7 +2,10 @@
' +
bolt-ai-dev-1 | ' return a + b;
' +
bolt-ai-dev-1 | ' }
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | " -console.log('Hello, World!');
" +
bolt-ai-dev-1 | " +console.log('Hello, Bolt!');
" +
bolt-ai-dev-1 | ' +
' +
bolt-ai-dev-1 | ' function greet() {
' +
bolt-ai-dev-1 | " - return 'Greetings!';
" +
bolt-ai-dev-1 | " + return 'Greetings!!';
" +
bolt-ai-dev-1 | ' }
' +
bolt-ai-dev-1 | ' +
' +
bolt-ai-dev-1 | " +console.log('The End');
" +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' // full file content here
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' Before providing a solution, BRIEFLY outline your implementation steps. This helps ensure systematic thinking and clear communication. Your planning should:
' +
bolt-ai-dev-1 | " - List concrete steps you'll take
" +
bolt-ai-dev-1 | ' - Identify key components needed
' +
bolt-ai-dev-1 | ' - Note potential challenges
' +
bolt-ai-dev-1 | ' - Be concise (2-4 lines maximum)
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' Example responses:
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' User: "Create a todo list app with local storage"
' +
bolt-ai-dev-1 | ` Assistant: "Sure. I'll start by:
` +
bolt-ai-dev-1 | ' 1. Set up Vite + React
' +
bolt-ai-dev-1 | ' 2. Create TodoList and TodoItem components
' +
bolt-ai-dev-1 | ' 3. Implement localStorage for persistence
' +
bolt-ai-dev-1 | ' 4. Add CRUD operations
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | " Let's start now.
" +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' [Rest of response...]"
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ` User: "Help debug why my API calls aren't working"
` +
bolt-ai-dev-1 | ' Assistant: "Great. My first steps will be:
' +
bolt-ai-dev-1 | ' 1. Check network requests
' +
bolt-ai-dev-1 | ' 2. Verify API endpoint format
' +
bolt-ai-dev-1 | ' 3. Examine error handling
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' [Rest of response...]"
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
@@ColeMedin And the part 3 : bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' Bolt creates a SINGLE, comprehensive artifact for each project. The artifact contains all necessary steps and components, including:
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' - Shell commands to run including dependencies to install using a package manager (NPM)
' +
bolt-ai-dev-1 | ' - Files to create and their contents
' +
bolt-ai-dev-1 | ' - Folders to create if necessary
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' 1. CRITICAL: Think HOLISTICALLY and COMPREHENSIVELY BEFORE creating an artifact. This means:
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' - Consider ALL relevant files in the project
' +
bolt-ai-dev-1 | ' - Review ALL previous file changes and user modifications (as shown in diffs, see diff_spec)
' +
bolt-ai-dev-1 | ' - Analyze the entire project context and dependencies
' +
bolt-ai-dev-1 | ' - Anticipate potential impacts on other parts of the system
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' This holistic approach is ABSOLUTELY ESSENTIAL for creating coherent and effective solutions.
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' 2. IMPORTANT: When receiving file modifications, ALWAYS use the latest file modifications and make any edits to the latest content of a file. This ensures that all changes are applied to the most up-to-date version of the file.
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' 3. The current working directory is `/home/project`.
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' 4. Wrap the content in opening and closing `` tags. These tags contain more specific `` elements.
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' 5. Add a title for the artifact to the `title` attribute of the opening ``.
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' 6. Add a unique identifier to the `id` attribute of the of the opening ``. For updates, reuse the prior identifier. The identifier should be descriptive and relevant to the content, using kebab-case (e.g., "example-code-snippet"). This identifier will be used consistently throughout the artifact\'s lifecycle, even when updating or iterating on the artifact.
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' 7. Use `` tags to define specific actions to perform.
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' 8. For each ``, add a type to the `type` attribute of the opening `` tag to specify the type of the action. Assign one of the following values to the `type` attribute:
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' - shell: For running shell commands.
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' - When Using `npx`, ALWAYS provide the `--yes` flag.
' +
bolt-ai-dev-1 | ' - When running multiple shell commands, use `&&` to run them sequentially.
' +
bolt-ai-dev-1 | ' - ULTRA IMPORTANT: Do NOT re-run a dev command if there is one that starts a dev server and new dependencies were installed or files updated! If a dev server has started already, assume that installing dependencies will be executed in a different process and will be picked up by the dev server.
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' - file: For writing new files or updating existing files. For each file add a `filePath` attribute to the opening `` tag to specify the file path. The content of the file artifact is the file contents. All file paths MUST BE relative to the current working directory.
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | " 9. The order of the actions is VERY IMPORTANT. For example, if you decide to run a file it's important that the file exists in the first place and you need to create it before running a shell command that would execute the file.
" +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' 10. ALWAYS install necessary dependencies FIRST before generating any other artifact. If that requires a `package.json` then you should create that first!
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' IMPORTANT: Add all required dependencies to the `package.json` already and try to avoid `npm i ` if possible!
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' 11. CRITICAL: Always provide the FULL, updated content of the artifact. This means:
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' - Include ALL code, even if parts are unchanged
' +
bolt-ai-dev-1 | ' - NEVER use placeholders like "// rest of the code remains the same..." or ""
' +
bolt-ai-dev-1 | ' - ALWAYS show the complete, up-to-date file contents when updating files
' +
bolt-ai-dev-1 | ' - Avoid any form of truncation or summarization
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' 12. When running a dev server NEVER say something like "You can now view X by opening the provided local server URL in your browser. The preview will be opened automatically or by the user manually!
' +
bolt-ai-dev-1 | '
' +
bolt-ai-dev-1 | ' 13. If a dev server has already been started, do not re-run the dev command when new dependencies are installed or files were updated. Assu'... 5141 more characters,
bolt-ai-dev-1 | messages: [ [Object] ],
bolt-ai-dev-1 | tools: undefined,
bolt-ai-dev-1 | tool_choice: undefined,
bolt-ai-dev-1 | stream: true
bolt-ai-dev-1 | },
bolt-ai-dev-1 | statusCode: 401,
bolt-ai-dev-1 | responseHeaders: {
bolt-ai-dev-1 | 'cf-cache-status': 'DYNAMIC',
bolt-ai-dev-1 | 'cf-ray': '8dd785a5b844e209-MRS',
bolt-ai-dev-1 | connection: 'keep-alive',
bolt-ai-dev-1 | 'content-length': '97',
bolt-ai-dev-1 | 'content-type': 'application/json',
bolt-ai-dev-1 | date: 'Mon, 04 Nov 2024 20:58:25 GMT',
bolt-ai-dev-1 | 'request-id': 'req_01J29jc9wfmTRXHxRMvZvvto',
bolt-ai-dev-1 | server: 'cloudflare',
bolt-ai-dev-1 | via: '1.1 google',
bolt-ai-dev-1 | 'x-robots-tag': 'none',
bolt-ai-dev-1 | 'x-should-retry': 'false'
bolt-ai-dev-1 | },
bolt-ai-dev-1 | responseBody: '{"type":"error","error":{"type":"authentication_error","message":"x-api-key header is required"}}',
bolt-ai-dev-1 | isRetryable: false,
bolt-ai-dev-1 | data: {
bolt-ai-dev-1 | type: 'error',
bolt-ai-dev-1 | error: {
bolt-ai-dev-1 | type: 'authentication_error',
bolt-ai-dev-1 | message: 'x-api-key header is required'
bolt-ai-dev-1 | }
bolt-ai-dev-1 | },
bolt-ai-dev-1 | [Symbol(vercel.ai.error)]: true,
bolt-ai-dev-1 | [Symbol(vercel.ai.error.AI_APICallError)]: true
bolt-ai-dev-1 | }
@@ColeMedin BTW, thank you for your response ! After several hours of struggle, I give up for today. I'll check your answer tomorrow and hope there will be a solution. PS : Thanks for your work, I don't know for me yet but I saw that a lot of people are very happy using this product.
can we add cofounder to this?
Good question
Yeah good question! I have still to look into Cofounder but I am going to this week! I know it's doable but I want to look into if it's best
@@ColeMedin awesome bro
Cofounder has Web Container support in the todo list! Just makes sense!
Awesome, I configured it on local machine with Docker but seems its not working when selecting ollama with qwen model, it seems to be taking Anthropic Model by default so ignoring Ollama.
Interesting... How could you tell it's ignoring Ollama?
@@ColeMedin I think that if you don't set up an API for Anthropic the tool won't work. I have the same problem (no anthropic pro account or credits for the API) and when I try to ask something I got the same error as everyone else. When I check the logs I got a message saying something about not being able to connect to Anthropic API. I also put an API but as I already mentioned, without credits, and the log message is different, saying something about "not enough credits on Anthropic". All these while choosing Ollama on the UI menu that you created for the tool.
Yeah I actually fixed this exactly problem recently I believe! There was an issue specifically in the container that I addressed.
Thanks man this stuff is awesome wish you luck, at some point I might add some of my web scraping with llm scripts into this , I dont do node js mostly python but you know I am sure their is a way this is gonna be the best 😃Pls upvote
great work, great project!
super kool.
As more contributions are added, how to update these updates in my install. I did the install via VS Code
Good question! All you have to do is run the git pull (either in the command line or in the Git extension for VSCode) and then restart the site/container to make sure the latest changes are reflected!
@@ColeMedin Could you please do a video on this?
@@ColeMedin thanks, will try this
@@JaMinaEngram I think a simple google search on git pull will help you execute it. I’m sure there are plenty of videos on TH-cam as well.
I have a question: I can only get this to jive with Ollama if I access it via localhost:5173. If I attempt to access it via the IP address (ie: 10.0.0.166:5173) I can access the frontend, but it will no longer display, or use, the Ollama models. This seems to be whether Ollama is hosted on the same machine as ottodev, or on another instance of ollama on the network (in this case, my ollama server). Both access methods require that I use the frontend at localhost, strictly, but I would prefer to run the frontend on its own server, if possible.
Any suggestions?
Thanks for the walkthrough on setting Ottodev up with Docker. Got it all up and running but it can't see any of my API keys. For each model where I've set the API key in .env I see the following error as per the OpenAI example here... "OpenAI API Key: Not set". Would appreciate some assistance - is the Discource community up and running yet?
Yes it is!
thinktank.ottomator.ai
Feel free to post there!
Did you make sure to rename the .env.example file to .env?
Awesome video!
I’m really excited to take on a challenge where I’ll develop an AI startup idea and launch it in just 48 hours starting on November 7. I want to see what I can create in such a short time! I’m still on the lookout for the perfect idea, so I’d love to hear any suggestions from you all.
I’m aiming for something fun, a bit challenging, but still manageable within 48 hours. Please share any ideas you have-I’ll choose one and dive right in. Thanks in advance for your input, and let’s see what unfolds!
Once I complete the AI project, I’ll share a video outlining the entire process I went through to create it. So buckle up-this journey is going to be exciting!
Thanks Suraj! Love what you're doing too! Suggestion: an AI agent to manage your Gmail/Outlook inbox.
Great idea, would prepare an analysis for it!
Thanks man!
The timer begins!
Done with it! Preparing to launch it.
Everything works good but except one thing, I entered all api key for all llm but since they are in free version i cant use them in docker, also, when sending a message to the only one working, it always repeat what the first message was saying, without seeing next messages
Could you clarify why you can't use them in Docker? What error are you seeing?
Can you integrate cline or claude computer into it or a seperate docker container somehow to edit actual files on your computer also
That would be a separate project really since Bolt.new is all about running in the browser, but that would be sweet!
Keep it up everyone!
As much as I would love to try this I am getting nothing but problems following the directions. I try the "npm run dockerbuild:prod" but get "ERROR: "docker buildx build" requires exactly 1 argument." So then I try the "docker build . --target bolt-ai-production" and then "docker-compose --profile production up" But I always end up with ".env.local: no such file or directory" Im wondering at what point does the .env.local get created.
I was able to get the dev environment to work but per the directions I am unable to get production to work.
I'm glad you got the dev environment working! What was the problem? Also lmk the error for the production build and I'd love to help!
even him dosent know what hes talking about, he says in the video to create .env file but then the bindings script for docker build is looking for .env.local .... : cat bindings.sh
#!/bin/bash
bindings=""
while IFS= read -r line || [ -n "$line" ]; do
if [[ ! "$line" =~ ^# ]] && [[ -n "$line" ]]; then
name=$(echo "$line" | cut -d '=' -f 1)
value=$(echo "$line" | cut -d '=' -f 2-)
value=$(echo $value | sed 's/^"\(.*\)"$/\1/')
bindings+="--binding ${name}=${value} "
fi
done < .env.local
bindings=$(echo $bindings | sed 's/[[:space:]]*$//')
echo $bindings .... in my case i cant even manage to make it work with the commands in the git readme file .... pitty ... save your time and run it locally : pnpm run build && pnpm run start
How much does it cost per message on average for the Claude 3.5 Sonnet API?
Great question! I believe you asked this in another video, but around ~$0.02 on average I would say. Given Claude 3.5 Sonnet is $3 per million input and $15 per million output tokens.
hey, a non dev here. Followed the docker set up but environment 'usually' does not build a preview in the ui giving a "You can now preview the updated homepage by running the development server." How do I rectify? Using gpt4omini (as have a bunch of tokens) - seems happy to chat but doesn't add much code.
That's weird! So sometimes everything works great but other times it doesn't? I'm guessing that is a limitation of GPT-4o-mini not executing the right commands versus it actually being an issue within the application. Unless you try with something like Claude 3.5 Sonnet and it still happens often!
@@ColeMedin thx. will investigate and give an update
I've got a big ask. Where do we start?? There so many places trying to say how to... Do you have any tips? I'm also looking at a new AI laptop? I'm all in!! Thanks a bunch
Great question - I love your enthusiasm! Are you asking where to start as far as contributing? That's really up to you and how you want to help improve the project! And I'll also be putting out a video soon on how to contribute and what that looks like!
I am not getting any preview of the development. 'Run command' kept running with 'npm install && npm run dev'. What to do?
same here
Sometimes it can take a while for the preview to show up... but if it isn't showing up even after a while, you can open the terminal with the button in the top right and then try to run the command yourself.
It's weird that sometimes the LLM will also not give the right command so that could be possible too.
@@ColeMedin tried running those command in console too, didn’t work.
Interesting... which model are you using?
Hi Cole! I just followed this video and installed everything and I'm using Gemini 1.5 Pro api. However, no matter what I enter in the prompt, I get a white screen in Preview. I'm also using Canary as you mentioned with no extensions installed. How can I troubleshoot this?
So that usually means there is an error with the LLM hallucinating a bad command or bad code. I cover troubleshooting steps in this video (timestamp included in link):
th-cam.com/video/5I9VgwauuzU/w-d-xo.html
I would participate in the project as a UX/UI designer if you need one 😊
I appreciate the offer! We definitely have a need for people like you to help with UI/UX!
This video says to use .env, they docker on Linux won't run unless the file is called .env.local. Even then it doesn't seem to read the Enviornment. OpanAI won't work.I am excited to be able to use this.
From what I understand Docker containers always use .env by default no matter the OS... did you test this yourself to be sure? I'm super curious, might have to update the docs!
Thanks Cole for the great work, I don't know where i have gone wrong but I keep getting this error, "There was an error processing your request" on the chat.
You are welcome! That error happens once and a while... what is the error message you see in the terminal where you started the site? I'd also check the developers tools (F12) in the browser and see the errors in the console there too.
hey man no matter what i do i just seem to keep on getting the same results and that is the " There was an error processing your request; no details were returned" . Please let me know what i am doing wrong as there is nothing more frustrating than trying multiple different routes and still just being caught up at the same error. thanks man i really do think what your doing is awesome by the way and i hope this project really takes off like it should.
Thank you! Sorry you are running into this error! What is the error message you see in the terminal where you started the site, or in the developer console in the browser? Typically that is much more helpful.
Is it just OpenAI key or other issues that keep me from seeing the preview of what it built? I'm only using OpenAI key.
That might mean the LLM decided to run the wrong commands... otherwise sometimes it can take a bit for the preview to load!
what is the meaning oTToDev name? does it come from "auto" dev?
Great question! It's a play on words to mean automated developer kind of as you said! The name also ties into something bigger behind the scenes I'm building for all of us!
whats the difference between this and openhands (prior OpenDevin)? is it the same purpose?
Those Devin like apps try to do a lot for you, usually way too much haha
oTToDev and Bolt.new is specifically meant to generate full stack web apps in the browser and it does that super well.
I follow exactly as you showed in this video. But I got this error message:
"There was an error processing your request"
There was API key setup problem.
I was using Ollama that runs on my computer.
I am not sure how to fix the error.
I have the same issue!
Sorry you're running into this! What is the error message you get in the terminal where you started the container? Usually that error message helps a lot more than the one on the frontend you referenced.
@@ColeMedin Buried in the text resulting from "npm run dockerbuild" is "| APICallError [AI_APICallError]: x-api-key header is required".
I've renamed the file .env with only this change to its contents:
# You only need this environment variable set if you want to use oLLAMA models
# EXAMPLE localhost:11434
OLLAMA_API_BASE_URL=localhost:11434
I'm on a Mac, if I go to localhost:11434 in Google Chrome Canary I get the message "Ollama is running" and the downloaded LLM I have is llama3.2
Can you see where the error is?
@ColeMedin Here is the error:
"APICallError [AI_APICallError]: x-api-key header is required"
**When I used OpenAI API key it works.**
**But having error when using Ollama API URL.**
My PC Setup:
- My PC OS is Windows 11.
- Installed Ubuntu OS for Windows.
- Then, installed Docker in Ubuntu and then Ollama with Openweb UI, and multiple llm models.
- It's working with no problem.
*Then I saw your video(This one) and followed your instructions*
- Clone the repo
- Installed Docker Desktop on Windows
and so on.
- It's working with OpenAI API Key.
-* But having problem with Ollama.*
@ColeMedin Any solution, please!
I'm having trouble getting the preview to work the run npm just stalls out cause the preview shares the same host port .. not sure I'm trying to claude troubleshoot this but im total noob
I was able to get docker and container set up fine bolt working great but no preview.. tried to fix it for hours last night
Might try the pinnocio method or just get the stax service for containers
Any help would be appreciated! Thanks bud!
Even if I run it with docker I can't get it working for preview either..
I think the issue is that I might have had a project open on my phone that was using the same port so im gonna try to work around that now and see if it fixes
Weird, and any code projects I generate on the bolt side, then I try to load it in stacks and it crashes the browser .. I wonder what's going on here
Weird... what is the error message you are getting?
@ColeMedin - i'm pretty new to this, but have managed to setup the oTTo/Bolt environment based on your instructions. Thanks. I'm also using it with Ollama. So cool. However, when I ask this local Bolt to create a someething (such as a basic 'hello world' html page) it prints out the code in the chat, but doesn't create the project files, run it in preview etc. on the right side of the screen. Any idea why this might not be running the way the original Bolt works. I must have done something wrong in my setup. Thanks
Actually, I've seen your latest video and this has answered my questions. The only problem now, is that my external API-Keys don't appear to be working. Not sure why as I have updated constants.ts and .env
Glad my latest video helped! That new error is strange... what is the error message you are getting?
@@ColeMedin it's something to do with my api setup. none of the api's i've setup (openrouter, openai, deepseek etc.) will work. my local Ollama works, just not the LLMs that I connect to via api key. whenever i try to communicate to the LLMs via the chatbox, a message pops up on the right of the screen "There was an error processing your request"
You can explain how can install in docker portainer
Greetings from colombia 🇨🇴
I cover that here!
th-cam.com/video/31ivQdydmGg/w-d-xo.html
Damn I'm getting following error on W10:
Docker Desktop requires Windows 10 Pro/Enterprise/Home version 19044 or above.
Can't seem to find .env.example file. Do I need to install VScode to see it?
It's at the base of the repo! You can also just run this without that and set your API keys within the UI now!
Just work for bolt at this point 😂 i dont mind giving 10$ monthly the only frustrating thing is token limit
Haha I want to stick with open source and they'd probably put me on their commercial version 😂
can u please add grok support to bolt so i can us my grok api ? would be much appreciated.
There is a pull request out for it now that I am going to review soon!
add feature to edit existing project
how to install locally in windows 10? i tried but not working
What error are you running into?
i got it to load in chrome canary but when i give it a prompt and hit enter it says "there was an error processing your request
"
Hmmm... what is the error message you see in the terminal where you ran the site? Or in the developer console in the browser?
I got it everything running but when I type make something I get an error there was an error pressing your request and it goes the same for all the api I add it looks like somehow the app can’t call the api
What is the error you are seeing in the terminal/developer console? Usually that is a lot more helpful!
I've got the local host opened in Canary. I can type, but there's no buttons.
You mean in the app preview?
i cant get the fork to work, all i getThere was an error processing your request and only openrouter works
Sorry you're running into that! What is the error you are seeing in the terminal where you ran the Docker container?
@@ColeMedinno docker. Just that error message when I try to use any model. Several ppl have reported this including in github. Please have a look
I'm taking a look, I believe I pushed a fix for Ollama that might help address this. Otherwise, what is the specific error message you are getting?
Why do I need google chrome canary? You mentioned that I need it, but you didn't say why I need it
Fair question! I'm not totally sure to be honest, even with the base open source Bolt.new you'll just get an error with Chrome saying it doesn't work with that browser and to use Canary. I believe it's something with the WebContainer that doesn't work with Chrome when you're running the app locally.
@good point, cheers
waitg I was runningi t in chrome and it was working, just not.....right, hmmm I should try another browser....thk u...
Yeah I'd try Chrome Canary or Firefox!
Will this ever became an app for Android, Or is that way to difficult?
At some point my plan is to host this somewhere for others to use and then it could certainly be an Android app!
Thanks for the reply yeah what's your working on is pretty awesome and groundbreaking The reason why I ask is because I like the mobility of the smartphone being able to work on it everywhere and I understand the power limitations compared to a laptop or a PC but of course these things are limitations at first.
great
I still need help with installing this. I have watched the video multiple times and read the notes back to back. I wonder if the problem is with Apple Silicon or me. Can anyone help me install this? I would be very grateful.
I'm sorry you're running into issues! What is the error you are encountering?
Cole are you an AI?
Second time someone has asked me that haha 😂what gives it away? lol
Also, please don't switch it up and start charging once you get to the juicey bits. I know we all have to make a living. But nothing worse than looking for a bootstrap, getting three quaters through a video and find out you have to pay. If its like a one off $49.99 or something, maybe. But I'm not looking to sign up per month. I admire your work, trust me. But please try and keep it for the poor in the cost of living crisis. Thats why we are here. 😉
My plan is to always keep this open source and free to use! I never want to start adding commercial features that you need to pay for like Bolt.new is doing, though I understand why they do that. That isn't for oTToDev though!
@@ColeMedin I was more referrring to when I signed up to your website, I got anxiety lol. No Im kinding, your'e good at what you do, just please, try and keep it low.
I gotcha! I'm never going to give up on always providing a ton of value for free, trust me! And oTToDev will always remain open source completely!
perfect x2
On a Mac using Docker and getting errors when I get to "Run with Docker" anyone else running into problems here specifically?
What errors are you getting?
make the user change the api keys in the settings
This is on the list of features to implement for sure!
why call it oTToDev??
I talk about it more in the video but it'll be good to give it a distinct name now that the project is evolving heavily instead of just calling it "the fork". Plus the name plays really well into the larger platform that I am working on behind the scenes for everyone!
add a plugin to vscode
Hi Cole, I am running into a ton of dependency errors when trying to run both on docker and on my OS locally. Despite over five attempts, I can't seem to get my config correct with the API keys. It either pulls the local instances or fails to map them correctly to the dropdowns. I can share what I have so far and debug further, I just don't have visibility if anyone else has this issue.
Please let me know how to best connect either 1:1 or in the discord, ideally before this coming sunday.
Sorry you're running into these issues! Is this for Ollama specifically? I actually made a fix yesterday for this that might help, so I'd try repulling the repo and running again!
@@ColeMedin Saw your new vid and doing so now. Thanks!
❤❤❤