How to Install Bolt new AI Locally : Complete Guide 2024

Introduction

This tutorial provides a comprehensive guide on how to install Bolt.new AI locally using Ollama models. You will learn each step, including troubleshooting common installation issues, to successfully set up and run the Bolt.new interface.

Step 1: Install Git

• Download Git from the official website: Git Downloads.
• Follow the installation prompts and ensure it is added to your system PATH.
• Verify the installation by opening a command prompt and typing:

  git --version
  

Step 2: Install Python

• Download Python from the official site: Python Downloads.
• During installation, check the box to add Python to your PATH.
• Confirm the installation by running:

  python --version
  

Step 3: Install Node.js

• Visit the Node.js website: Node.js Downloads.
• Download the version suitable for your system and complete the installation.
• Check your installation with:

  node --version
  

Step 4: Install Visual Studio Code

• Download Visual Studio Code from: VS Code Downloads.
• Install it by following the on-screen instructions.

Step 5: Install Ollama

• Go to the Ollama website: Ollama.
• Follow the instructions to install Ollama on your system.

Step 6: Install Ollama's Model

• Use the following command to install the desired model:

  ollama run qwen2.5-coder
  

Step 7: Set the Execution Policy

• Open PowerShell as an administrator and run:

  Set-ExecutionPolicy RemoteSigned
  

Step 8: Clone the Bolt.new Repository

• In your command prompt, execute:

  git clone https://github.com/coleam00/bolt.new-any-llm.git
  

• Navigate to the cloned directory:

  cd bolt.new-any-llm
  

Step 9: Install npm

• If npm was not installed with Node.js, run:

  npm install
  

Step 10: Install pnpm

• Install pnpm globally using:

  npm install -g pnpm
  

Step 11: Configure the .env.example File

• Rename the .env.example file to .env.
• Open the file in VS Code and configure the necessary environment variables as needed.

Step 12: Create Modelfile

• In the root directory of Bolt.new, create a file named Modelfile.
• Add the following content:

  FROM qwen2.5-coder
  PARAMETER num_ctx 32768
  

Step 13: Create a Custom Ollama Model

• Run the following command to create your custom model:

  ollama create -f Modelfile qwen2.5-coder-bolt
  

Step 14: Run Bolt.New UI

• Start the Bolt.new interface by executing:

  pnpm run dev
  

Step 15: Troubleshooting Common Issues

• Deleting an Ollama Model: Use the command:

  ollama rm [exact model name]
  

• Ollama Model Not Visible: Ensure that the Ollama service is running and check the localhost address:

  http://127.0.0.1:11434
  

Conclusion

You have now successfully installed and set up Bolt.new AI locally with Ollama models. If you encounter any issues, refer back to the troubleshooting steps above. As a next step, you can start exploring the functionalities of Bolt.new and consider experimenting with other models or features. Happy coding!