Uploaded August 2026 | Updated November 2026, 1 day ago
In this episode of Coding with AI, I build a little application I’m calling "Pet Llama" — a simple front end for a local Ollama model running entirely on my Mac.
The idea started because I wanted AI to review some text that I didn’t want to send to an online AI service. Ollama worked, but interacting with it directly was clunky. So naturally, instead of just living with that, I decided to build my own user interface.
I use ChatGPT to help design the idea, then hand a prompt to Cursor and watch it build the application. Along the way we:
* Talk to a local Ollama model
* Stream the model’s “thinking” and final response into the browser
* Build reusable skills/prompts
* Create a “Polish” skill for improving text
* Add a way to stop a runaway Ollama request
* Watch Cursor test its own code using its agentic browser
* Discover that building an AI tool *with AI* can get hilariously meta
This is not a polished tutorial. It is a real-time look at how I use AI-assisted programming: talking through an idea, refining the requirements, watching the tools make mistakes, fixing them, and ending up with a working MVP.
The source code for "Pet Llama" will be on my GitHub:
github.com/csev/pet-llama
More Coding with AI from Dr. Chuck:
dr-chuck.com
In this episode of Coding with AI, I build a little application I’m calling "Pet Llama" — a simple front end for a local Ollama model running entirely on my Mac.
The idea started because I wanted AI to review some text that I didn’t want to send to an online AI service. Ollama worked, but interacting with it directly was clunky. So naturally, instead of just living with that, I decided to build my own user interface.
I use ChatGPT to help design the idea, then hand a prompt to Cursor and watch it build the application. Along the way we:
* Talk to a local Ollama model
* Stream the model’s “thinking” and final response into the browser
* Build reusable skills/prompts
* Create a “Polish” skill for improving text
* Add a way to stop a runaway Ollama request
* Watch Cursor test its own code using its agentic browser
* Discover that building an AI tool *with AI* can get hilariously meta
This is not a polished tutorial. It is a real-time look at how I use AI-assisted programming: talking through an idea, refining the requirements, watching the tools make mistakes, fixing them, and ending up with a working MVP.
The source code for "Pet Llama" will be on my GitHub:
github.com/csev/pet-llama
More Coding with AI from Dr. Chuck:
dr-chuck.com










