This repository is a simple combination of Ollama, Open WebUI and Automatic1111-Stable Diffusion models that works together in a docker compose file, ready to be deployed in a on prem environment.

You would need a computer with quite some Memory resources 16GB+, it is setup to work on only the CPU but GPU is possible to enable by providing the nvidia/amd runtime in the compose file. You would also need docker and docker-compose installed on the computer.

1. Clone this repo: git clone https://github.com/jolin1337/kvadratgpt.git && cd kvadratgpt
2. If you want GPU support go into ./docker-compose.yaml and ad runtime: nvidia to the services you'd like to have GPU support for, otherwise you are already setup to start the application
3. Now just run: docker compose up -d automatic1111 open-webui ollama

Current supported features are constantly updated by the creators of these amazin products, although this is a list of some major ones:

• ChatGPT like Interface (open-webui)
• ChatGPT like Custom Agents (open-webui)
• Information retireval through RAG (open-webui)
• Image generation based on a prompt (automatic1111)
• Community driven GPT Model and prompt support (open-webui community)
• All is local and no information is shared with anyone but the users of the local instance
• Integration support with other tools such as (but not limited to) developer IDE:s Continue and ollama