De-Googling my life (to the extent that I could)

Rob de Wit
9 min readMay 30, 2024

--

Photo by La-Rel Easter on Unsplash

About a year ago, I bought a NAS to organise my photo library. I was tired of keeping three external HDDs in sync, and a NAS seemed like a great solution: built-in redundancy, automated backups, and I didn’t need to carry it with me.

While I’m still pleased with my NAS, it proved to be a bit of a gateway drug. It already stores all of my photos. Why not migrate away from Google Photos and host https://photos.personaldomain.tld to share albums with friends? And if I can use my NAS to host documents as well, I might as well set up a Google Drive alternative…

All of this coincided with the steady decrease in the quality of Google Search. These past few weeks, forums like Lemmy have been rife with terrible AI results advising you to leave your dog in a hot car, eat a small rock a day, or drink urine to pass kidney stones. But even before then, Google’s results seemed to be getting worse and worse.

Cue me switching over to Kagi for my searches. It’s giving me the results I want, and I pay for them with money instead of my data.

This got me thinking: how far can I take my de-googling? As someone who’s been embedded in their ecosystem for well over a decade, I relied on Google services in more ways than I had thought. Perhaps I should’ve switched the moment they dropped their “do no evil” motto, but better late than never.

Here’s how I de-googled my life — to the extent that I could. I still watch YouTube on a daily basis, and I sometimes look things up on shopping.google.com. But progress is progress.

This post focuses on the services I ended up picking. Behind the scenes, I heavily use Docker, Tailscale for networking, and Terraform and PyInfra for my reverse proxy. I’ll probably follow up with a post dedicated to my infra repository.

Search: Kagi

As outlined above, I’m quite happy with Kagi. Its results are better than Google’s and I prefer to pay with money rather than with my personal data.

  • Cost: $9 per month
  • Compared to Google: great
  • Alternatives I considered: DuckDuckGo

E-mail: Proton

E-mail was the one service I did not want to self-host under any circumstances. My entire life is tied up to my personal domain, and I don’t trust my SysOps skills enough to guarantee 99.999% uptime on my NAS. Any e-mails I send would likely also end in the spam directory.

Proton is committed to privacy, and I like that they’re based in Europe. While they have received some criticism for not being completely watertight, this is hardly a factor in my threat model. And I find that the company is quite transparent about its actions.

The migration to Proton vindicated my use of a personal email domain. Switching over took a few updated DNS records and an import of my old e-mails. I didn’t have to update my accounts anywhere. Even if you stick to Gmail, I highly recommend getting a custom e-mail domain to prevent lock-in.

I am currently using Proton’s business plan, primarily because I use more than three e-mail domains.

  • Cost: starts with a free plan, business is €11 per month
  • Compared to Google: good
  • Alternatives I considered: Tuta Mail

DNS: Mullvad (through Tailscale)

My DNS switch to Mullvad might be my favourite change in this list. As I described a year ago, I use Tailscale to connect my personal devices to my NAS. Tailscale partners with Mullvad to provide exit nodes for your network, letting you browse the web with a different external IP address. This add-on to Tailscale costs $5 a month.

A completely free service Mullvad provides is Mullvad DNS, which you can use regardless of whether you use their VPN. I configured Tailscale to use Mullvad’s nameservers.

Whenever I request a web page, my computer will ask Mullvad to resolve the request. The cool thing is that Mullvad can automatically block certain content. I use base.dns.mullvad.net, which automatically blocks all advertisements, trackers, and malware across all of my devices.

  • Cost: free
  • Compared to Google: great
  • Alternatives I considered: none

Calendar: self-hosted (Synology)

I’ll be honest… This is where I really started to feel the friction. I have several shared calendars with friends and my partner. This works perfectly when everyone is using a Google account. Giving your partner write access to the shared calendar is trivial.

This is not the case when you’re using a different calendar provider. I tried Proton’s calendar, but its experience with third-party clients was awful. Proton insists on end-to-end encryption, which is a good principle, but it also means you can only use their client. And their client doesn’t allow you to edit events from different accounts.

I want my work and private calendars in the same client (Fantasical), both on my laptop and phone. That’s why I ended up settling for a self-hosted Synology Calendar instance. It appears to be based on Radicale, which I might yet switch to because Synology’s implementation has a few quirks.

I have set up https://calendar.personaldomain.tld to share with my partner. Because we don’t have accounts on the same server, we cannot edit the same calendar. Our workaround is two calendars: I can edit Rob + Alma calendar, and she can edit Alma + Rob calendar. The other person just has view access.

  • Cost: free (if you have a Synology NAS)
  • Compared to Google: poor
  • Alternatives I considered: Radicale, Baikal, Fruux, NextCloud

Contacts: self-hosted (Synology) + Apple + Proton

This one goes hand-in-hand with the Calendar, because it runs on the same CalDav server. That’s how I settled for Synology Contacts.

In practice, however, Proton keeps track of familiar e-mail addresses. And every time I want to call someone, I look them up in my iPhone’s contacts. I’ll admit that these are probably not entirely in sync, so I have some room for improvement.

  • Cost: free (if you have a Synology NAS)
  • Compared to Google: okay
  • Alternatives I considered: it’s a mess already

Photos: self-hosted (Synology) + Apple

The one that started it all: how do I self-host my photos? I upload all of the pictures I take with my dedicated camera to the Raw photos directory on my NAS. Once I sort them out and edit them, I add the exports to the Photos directory.

It’s this Photos directory that Synology Photos picks up. This gives me a nice timeline overview, as well as some more advanced features like facial recognition. I’ve set up https://photos.personaldomain.tld for sharing photos with friends and family.

There’s a bit of data duplication going on with Apple here. I upload my exports to Apple Photos to have them available in two places. The Photos directory on my NAS also contains Mobile backups, to which Synology’s iPhone app automatically uploads all of my mobile pictures.

  • Cost: free (if you have a Synology NAS)
  • Compared to Google: good
  • Alternatives I considered: Ente, Immich

Drive: self-hosted (Synology + Paperless-ngx)

If you heavily use Google Drive to collaborate on documents, this one will be much trickier to replace. I use Drive mostly for storing documents and occasionally creating a spreadsheet or presentation.

That’s why Synology Drive is good enough for me as an alternative. The entire purpose of my NAS is to store files, so it makes sense to have a tool that can access that very filesystem.

I can type the occasional document on Drive, and I’ll likely switch to an alternative for more intensive workloads. I write letters in Overleaf, for example, blog posts in Markdown, and create presentations in Keynote, Figma, or a Jupyter Notebook. The resulting files I can then store on my NAS again.

At work, where I do collaborate much more frequently, Google Drive has largely been replaced with Notion — which is more flexible than Drive.

One extra tool I would like to highlight in this category is Paperless-ngx. When I mentioned document storage, I mostly referred to documents that are natively digital. With Paperless-ngx, I now also have a digital archive of my physical correspondence (mostly tax letters). With a press of the button, my scanner scans directly to the NAS, where Paperless-ngx indexes the file and makes it searchable through OCR.

  • Cost: free (if you have a Synology NAS)
  • Compared to Google: okay
  • Alternatives I considered: NextCloud

https://paperless-ngx.com

  • Cost: free
  • Compared to Google: great
  • Alternatives I considered: Docspell

Maps: Apple Maps

Fine, there’s a reason this one is near the end of the list. I haven’t found a satisfying replacement for Maps yet. I know OpenStreetMap exists, and I’m glad it does, but it just doesn’t cut it in my daily life.

Apple Maps is an obvious choice because I’m already using an iPhone anyway. Anno 2024 it’s a very usable product, although it’s still a bit of a downgrade compared to Google. Apple’s data on public places like restaurants is less extensive, making it harder to find what you’re looking for.

It helps that Apple recently released support for cycling navigation here in the Netherlands — which took them surprisingly long.

  • Cost: free
  • Compared to Google: okay
  • Alternatives I considered: none

Identity provider: Auth0

The final migration on my list is one I hadn’t anticipated beforehand. I avoid Log in with Google/Microsoft/Facebook buttons wherever possible. All my accounts are tied to an e-mail address, and all my passwords/passkeys are stored in 1Password. This works perfectly well and prevents lock-in.

“All of your accounts?” you ask? Well, there’s one critical service in my stack that requires a dedicated identity provider. Tailscale doesn’t want to store your e-mail and password — so a year ago I signed up with Google.

I looked into self-hosting an OIDC service, but again, I didn’t trust my SysOps skills enough for a critical piece of infrastructure. If I cannot log in to Tailscale, I cannot access most of the services I listed above. That’s why I opted for Auth0 instead.

It’s a very extensive product, and Okta would probably have been a more appropriate solution. But because Auth0 has a free plan and Okta doesn’t, I now have the rainforest and gorilla alongside the banana I wanted.

  • Cost: starts with a free plan
  • Compared to Google: great
  • Alternatives I considered: Authelia, GitLab, GitHub, Apple

And for the rest…

For completeness’ sake, here’s a quick list of software I was already using instead of Google’s products:

  • Phone: Apple iPhone
  • Browser: Arc (even though it’s still Chromium-based) and Firefox
  • Password manager: 1Password
  • Notes: Apple Notes
  • Reminders: Apple Reminders

Conclusions

I’ll be honest: de-googling is not a process I would generally recommend. There’s a reason walled gardens are largely accepted by users: if one company hosts your calendar, e-mail, photos, and maps, odds are that all services play together nicely.

Moreover, network effects are difficult to overcome. I don’t think I’ll cancel my YouTube Premium subscription anytime soon.

Even if you manage to jump the wall, you might just land right in another garden. Apple Maps is a decent alternative to Google Maps, sure. But you’re facing many of the same principle issues.

That said, it is really satisfying to be in charge of your own data. Getting esoteric Docker images to run can be challenging and frustrating, but it’s really cool when they finally work.

If you want tools that work straight out of the box, you’ll likely end up paying for a few monthly subscriptions. I can definitely recommend this for services like Tailscale, Mullvad, and Proton. But it is a privileged position, and I recognise that not everyone can afford that — similar to how the cost of a NAS will be prohibitive to many.

Luckily, free and open software provides some hope here. Many open-source tools are as good if not better than Google’s services. Projects like Radicale, Paperless-ngx, and Immich are essential to ensure we don’t become trapped entirely in big tech ecosystems.

So: support the FOSS communities you care about.

In Dutch: “Your Google Workspace Business Standard-subscription has been cancelled”, as of today.

--

--

Rob de Wit
Rob de Wit

Written by Rob de Wit

I'm an engineer who likes working with data to help people solve problems.

No responses yet