This project leverages open source tools to enable "vibe coding," an innovative approach where developers collaborate with AI to generate code from natural language prompts. It integrates AI assistants and models, transforming the coding process into a seamless, interactive, and creative partnership for faster software development.
-
nvidia-rtx-ubuntu-setup
https://github.com/rootveda/nvidia-rtx-ubuntu-setup -
Ubuntu-ai-lmstudio
https://github.com/rootveda/ubuntu-ai-lmstudio-setup
Launch lmstudio:
./lm-studio
Switch to Developer mode by clicking the 'Green Terminal Icon'
Click on Select a model to load
Select the default model (e.g., "Gemma-3") or another downloaded model
Once the model status is READY:
Increase Context Length from the default 4096 to 20000
(set per GPU, minimum length should be 15000)
Click on Reload to apply changes.
Toggle the status button from stopped to start (or press
Ctrl+R).
Take note of the server address & llm name:
http://127.0.0.1:1234
google/gemma-3-12b
Cline is an AI development assistant which integrates with Microsoft Visual Studio Code.
-
Launch Visual Studio Code.
-
Open the command palette (
Ctrl+P), paste the following command and press Enter:
ext install saoudrizwan.claude-dev
-
Click on Cline from the sidebar.
-
SignIn if prompted.
This will take you to app.cline.bot to create your account. -
Click on the settings icon and configure:
- API Provider:
LM Studio
- Check Use custom base URL and enter:
http://127.0.0.1:1234 - Model ID:
google/gemma-3-12b
-
Click Done.
-
You’re ready to start building! Copy and paste this prompt into the Cline chat window:
Hey Cline! Could you help me create a new project folder called "hello-world" in my Cline directory and make a simple webpage that says "Hello World" in big blue text?
For additional installation steps or troubleshooting, check the official Cline documentation:
https://docs.cline.bot/getting-started/installing-cline
Enjoy your AI-assisted vibe coding experience with lmstudio-vscode-cline!