FreeGPT-2 is born from the merger of two powerful software: Ollama and Open Web UI. Ollama, known for running large language models locally in CLI, is paired with the user-friendly interface of Open WebUI. This combination provides a clean and intuitive experience, making it an invaluable tool for working with locally hosted language models in a sovereign fashion. This repository creates the s9pk package that is installed to run FreeGPT-2 on StartOS. We gratefully acknowledge the use of code from both Ollama and Open WebUI, which has made the creation of FreeGPT-2 possible.
Prior to building the free-gpt package, it's essential to configure your build environment for StartOS services. You can find instructions on how to set up the appropriate build environment in the Developer Docs.
Clone the FreeGPT-2 package repository locally.
git clone https://github.com/Start9Labs/freegpt2-startos.git
cd freegpt2-startos
To build the FreeGPT-2 service as a universal package, run the following command:
make
Alternatively the package can be built for individual architectures by specifying the architecture as follows:
# for amd64
make x86
or
# for arm64
make arm
Before installation, define host: https://server-name.local in your ~/.embassy/config.yaml config file then run the following commands to determine successful install:
ℹ️ Change server-name.local to your Start9 server address
start-cli auth login
#Enter your StartOS password
make install
Tip: You can also install the free-gpt.s9pk by sideloading it under the StartOS > System > Sideload a Service section.
Go to your StartOS Services page, select FreeGPT-2 and start the service.
Done!