# MattGPT

That's right everyone, I have set up a LLM on my server that is accessible to select matt-cloud users. Remember you can check your group membership at the [groups site ](https://groups.matt-cloud.com/ "groups site")- look for **mattgpt-access**. MattGPT is located at [https://gpt.matt-cloud.com](https://gpt.matt-cloud.com). The source code for this can be found [here](https://github.com/open-webui/open-webui).

When you get there, you can log in with the trusty **Continue with Matt-Cloud** button.

[![image.png](https://kb.matt-cloud.com/uploads/images/gallery/2025-09/scaled-1680-/image.png)](https://kb.matt-cloud.com/uploads/images/gallery/2025-09/image.png)

When you login, you will be greeted with the default model of **deepseek-coder-v2** because I usually use LLM's as a computer science intern. In the top corner you can change between models. If there's a model you'd like to try, you can search for it [here](https://ollama.com/library) and send Matt an email or text or whatever and ask if it can be installed. Answer is much more likely to be yes than no assuming the system is operating properly.

[![image.png](https://kb.matt-cloud.com/uploads/images/gallery/2025-09/scaled-1680-/Uytimage.png)](https://kb.matt-cloud.com/uploads/images/gallery/2025-09/Uytimage.png)

I currently have 12 models installed here. Default is deepseek-coder-v2 because it's a decent computer science intern, but I believe you can configure your own default model.

<table border="1" id="bkmrk-phi4%3A14b-qwen2.5-cod" style="border-collapse: collapse; width: 74.7619%; border-width: 0px; border-style: none;"><colgroup><col style="width: 34.6097%;"></col><col style="width: 65.3916%;"></col></colgroup><tbody><tr><td style="border-width: 0px;">- [phi4:14b](https://ollama.com/library/phi4)
- [qwen2.5-coder:14b](https://ollama.com/library/qwen2.5-coder)
- [gpt-oss:latest](https://ollama.com/library/gpt-oss)
- [qwen2.5:latest](https://ollama.com/library/qwen2.5)
- [llama3.2:latest](https://ollama.com/library/llama3.2)
- [dolphin-mistral:latest](https://ollama.com/library/dolphin-mistral)

</td><td style="border-width: 0px;">- [gemma2:latest](https://ollama.com/library/gemma2)
- [deepseek-r1:latest](https://ollama.com/library/deepseek-r1)
- [deepseek-coder-v2:latest](https://ollama.com/library/deepseek-coder-v2)
- [deepcoder:latest](https://ollama.com/library/deepcoder)
- [qwen2.5vl:latest](https://ollama.com/library/qwen2.5vl)
- [gemma3:12b](https://ollama.com/library/gemma3)

</td></tr></tbody></table>

This is actually a pretty advanced system and I don't know exactly what its full capabilities are, so if you find something powerful or scary be gentle and let me know. This is AI here, and while these are "just LLM's" if those LLMs have certain access or information they can get spicy.