Ollama is an amazing tool to run AI models locally
on your laptop. On Mac, Windows, and Linux it's so incredibly easy to install and get up and running,
far easier than any other tool for this purpose.

But often in the Discord we see folks ask how
to update their Ollama installation. So in this video, I want to show you everything you need to
know to update your Ollama instances to both the current version as well as to a prerelease or
any past release since it's the same process.

When we first started building Ollama the install
was a bit rough, but we got it in line within a couple of weeks and it's been great ever since.
I say we because I was on the Ollama team when

we started working on it. I am no longer on
the team, instead focusing on building out this YouTube channel, but we are all still friends
and I still talk to the team regularly. This video is part of my weekly series called the Ollama
Course. If it's the first video you have seen,

take a look at the playlist to see
the rest of the videos in this course. I'll look at this update process from the
perspective of a Mac user, Windows user, and Linux user, as well as a Docker user on
Windows and Linux in that order. If you are

using Ollama via WSL on Windows, you really
should transition to using the native Windows install since it is so much faster. And if you
are using Docker on Mac, then you should also switch to the native install since Docker on Mac
has no access to the GPU. If you only care about

one of those platforms, the timestamps to zoom
ahead to that platform should be onscreen now. So Mac first. Now there are a number of
ways to install Ollama, but only one way that is recommended by the team. And that is the
official installer on the Ollama home page. Some

folks like to install with Brew, but there
isn't really any benefit to doing so. The Brew package is not maintained by the authors
of Ollama. Some say it's easier to upgrade, and that may have been true for 2 weeks or
so back in June of 2023, but since then it's

just easier to use the official installer, which
is also the most stable way to install Ollama. So I am assuming that you have done that. Now when
there is an update to Ollama, you will see the menu bar icon change to show a little arrow on it.
Click the icon, and choose "Restart to Update".

It's possible that the text shown will change
at some point in the future, but it should be something like that. Ollama will update itself and
be back up and running for you to use. And that's it, nothing else for you to do. How about if you
want to update to a new prerelease version? Well

just go to the Ollama homepage and then click
on the link to Github. Then on the right side, click on 'Releases'. At the top there is the
latest version. If there is a prerelease, then that will be at the top, otherwise it's
the latest full release. A prerelease is just a

version that is still not fully fleshed out and
may have some issues. You shouldn't update to this version unless you need a specific feature
it promises to solve. If you want to go back to an earlier version, then scroll down on the
releases page to find whichever version you

want. Expand assets. Then you will want to download the file called ollama-darwin.zip. This is the full installer and is the file you most probably want. Make sure you click the Ollama icon in the menu bar and choose quit, then unzip ollama-darwin and run the

ollama app that is extracted. Since the app is probably not in the correct directory, it will prompt to move it to the right place. After that has happened, you should be all set. So that was pretty easy. You know, it's almost as easy as giving a like and subscribe

to this video. That makes such a huge difference to the channel and I greatly appreciate everyone who does it. Thanks for being here. Let's move on to Windows. As with the Mac, there are different alternate ways to install Ollama,

but most of the time you really should be using the default installer which you can find on the Ollama home page. Once Ollama is installed, it will alert you when there is a new version to install. So click the ollama icon on the taskbar and then click "Restart to Update".

Ollama will download the latest version and update itself for you. How about if you want to install something other than the latest released version...maybe it's a prerelease or even an earlier version. Well, go to Ollama.com, then click on GitHub. On

the right is releases and you can see a list of all the available releases, with the latest one at the top. Scroll up or down to find the prerelease you want or an older version. Expand Assets and find OllamaSetup.exe.

Download that, then make sure to quit your running
ollama if it is running right now. Run OllamaSetup
and walk through the prompts. And you are now all
set on your chosen version of Ollama on Windows.

OK, on to Linux. Just like we saw with Mac
and Windows, there are a few different ways to install Ollama on Linux. The best is the
official install script. There are other install packages available but they sometimes
default to prereleases and cause problems.

To update Ollama to the latest version on Linux
you need to do a different process than the
other platforms. Since there is no menubar or
taskbar icon to click, you need to simply run
the install script again. So that's `curl -fsSL
https://ollama.com/install.sh | sh`. It will get

up and running super quickly. If you want to see
what that script is doing, simply take a look at
the script at the URL and you can see it's pretty
simple. If you want to install a different version
such as a prerelease or older version, just set
the Ollama version environment variable at the

command line before running the shell. So `curl
-fsSL https://ollama.com/install.sh` then pipe
that to `Ollama_version=0.4.0` or whatever
version you like, then `sh`. It really is that easy.

The final platform to look at is Docker. There
is nothing specific about Ollama when updating
it using Docker. It's the same process for any
Docker-based application. Stop the container,
remove the image, pull the latest version
from Docker Hub and then run it again using

the same command you did before. Because
you mounted a volume for the models, they are intact when deleting the image and
starting again. So that's easy, but even easier is to use a second container called
Watchtower. Watchtower runs at a regular interval

and will then update all the containers you want
it to update. Here is the docker-compose file I use for watchtower. It runs every few hours and
keeps things up to date. I do this on a server on which I run n8n and searxng and nocodb and I never
have to think about updating. It's pretty amazing.

And that is pretty much everything you need to
know about updating Ollama on every supported platform. It's pretty painless regardless of
the platform you use. I hope this was useful to you and that you enjoyed the video.
Thanks so much for watching. Goodbye.

Let's Update Ollama Everywhere

I was a founding maintainer of Ollama, the first evangelist at Datadog, and a former organizer of DevOps Days Seattle, DevOps Days Boston, and Serverless Days Boston. In my day job, I tour the country speaking at conferences and writing about the company I work at. But I am also passionate about gadgets. Some folks don't think that word is appropriate, but it totally fits here. If you meet me in person you will see me light up when I can share with you my latest gadget, tool, or utility. And this is where I get to share that with all of you. You can find more about me at my website (http://technovangelist.com) or on Twitter (http://twitter.com/technovangelist)