Integrate the "AI chatbot" feature with local LLMs and frameworks, e.g. Ollama, WebUI, etc.

[ntsarb]

Hello,

I noticed the "AI chatbot" feature in Firefox, which provides access to public cloud AI chatbots. However, it is now feasible to install a Large Language Model locally, e.g. using Ollama and other frameworks (e.g. WebUI), either on the same computer or on another computer at home (same LAN).

Wouldn't it be great to provide an integration feature of Firefox with our own LLM framework?

Regards

[Picpic]

Hello,
I confirm, it would be super nice to be able to connect to an ollama server 🙂

[devurandom]

See-also: https://connect.mozilla.org/t5/ideas/ai-ollama-integration/idi-p/85710

[sf1tzp]

There is an awesome suggestion in that thread ^

If you have Open WebUI (or SillyTavern or something) running, you can just put the URL into browser.ml.chat.provider in about:config.

The sidebar just loads and displays a webpage (ChatGPT or whatever) anyway.

---

You need to run a UI for Ollama at the moment but that feels like a reasonable compromise. It works for me at least.

[robertwilliam_0]

I think local LLM integration would be a valuable addition to Firefox’s AI chatbot capabilities. Tools such as Ollama make it possible to run models locally, which could give users more control over their data and reduce the need to send sensitive prompts to cloud-based services.

It would be interesting if Firefox could connect to a locally running LLM through a configurable API endpoint, allowing users to choose their preferred model or framework. This could also make the AI chatbot useful in environments where privacy, offline access, or control over the underlying model is important.

A flexible integration layer supporting local and cloud-based models could give users much more choice while keeping privacy as a key advantage.