I just spent the past 2 weekends migrating from Lightroom CC with just under 1TB of photos, mostly RAWs, making up about 50,000 files. I kind of made it up as I went and I spent a lot of time testing and validating. I'm going to do a writeup here to hepefully help ayone else coming acrross this and/or in case someone has ideas.

My Requirements:

1. Move all photos off Lightroom and preserve both RAW files and snapshot renders at high quality. I intend on staying on Lightroom because it's my editor of choice. I've litigated Darktable in comments on Lemmy, and while I love the software and the concept, I use Lightroom for my semi-professional work
2. (The hard part) Metadata. I absolutely need my album organization, my picks, and my star ratings from Lightroom. More on this below
3. All data must move to my self-hosted Immich with a TrueNAS pool behind it. The stack is Proxmox running a Truenas VM and a Debain VM. The Debian VM runs docker where Immich runs. Connectivity is driven by Netbird mesh VPN. That is a limitation for sharing and I'm ok with it.

Getting your pictures out

• Adobe doesn't make this easy just below the surface. There are open source tools intended to help one migrate from Lighroom Classic catalogs to Immich, but CC is very different. The export options are very particular.
  • Adobe lets you export a massive dump of pictures and renders as a real export option. It's through their website and it will email you when a dump is ready. Said dump will contain multiple zip files each with RAWs and renders. NO Metada or structure is included and NO edit data.
• You can "Archive" from Lightroom CC (tested on Mac) and that will output originals + the edit file AND it will preserve the root album from which your photo came. But no metadata beyond that (no ratings, folder structure, picks). The bad part is that this deletes your files once it's done. You can still restore fromTrash but this is definitely a hostile action against users.
• You can just output EVERYTHING. This is for the truly insane as you sit there and wait for Lightroom to render each one of your pictures individually. If you do original + edits, that returns the same thing as Archive but without structure.
• The secret 4th option: you can (as of this writing) parse Lightroom's internal catalog and get Albums, picks, star ratings, and folder structure. This was a massive unlock for me. This file was at ~/Library/Application Support/Adobe/Lightroom/

My solution going forward

Because I was discovering this as I went, I tried some version of all of the options above, but going forward, I have a process which should be much easier.

First, I will parse the catalog which is a SQLite db and extract the newest "version" of each picture and the content I care about. I will store it in a CSV for long-term holding.

Once metadata is stored, I'll go and request the archive through the website. That will take a few hours to possibly days (it took me about 2 hours on a Friday night). Once that's done, you get an email and a link to a personalized site. You have to parse the (for me) 50+ links and download them sequentially. It took me a few hours to download 1 TB limited by my local bandwidth (for which I'm eternally grateful). With my TrueNAS storage, I have the luxury of landing the data and uploading it to Immich separately.

For XMPs, the edit data, you have to get that from a local source so either export originals or the desktop app "Archive All" option.

You'll have to decide what you need and maybe you have fewer requirements than me.

Landing the Pictures

The next challenge is grabbing all of that data, pictures and ratings and applyhing them in Immich. Luckily, it has a fantastic API so I can script the whole thing. Most of the mapping happened by picture name and there were a few times where I had to dig into exif to identify a picture. Overall, i think I had 40 or so files out of 50k+ which couldn't be mapped.

Encountered Problems:

• As mentioned before, this was a try-it-until-it-works situation for me. I had to download all of my pictures 2 times by 2 different methods (I didn't document this but there was an SMB share involved).
• Stars are a new object and I had to manually enable it in Immich
• I couldn't replicated nested album functionality in Immich and chose to keep the root album name + a tag of the folder structure. That worked well for me
• Lightroom tags could not be exported. I don't use them a lot but that could be a blocker for some.

Disclaimer

• Every word in this write up was written by me on my keyboard
• I used a self-hosted local AI to do this work.
• I know enough to be in control even at a technical level, but the time savings on this project with AI really made it possible
• The AI found the catalog and parsed the data out of it. It would not have occurred to me.

Future

• I have all of this work on my personal Forgejo and if there's enough interest, I may post publicly, though this write-up should serve as a starting point for anyone who wants to replicate my process

Yeah Google does the same. When you Takeout your photos they split the images and metadata into separate files. Bunch dicks.

I haven't documented this yet but I did do Google Takeout. It is just as bad and they rotate credentials every few minutes so they make it really hard to download the 80 files they output my pictures into

Immich has built-in compatible tooling to handle Takeout, bless them

Takeout

Edit: voyage0 correctly pointed out that I used/meant immich-go, not built-in tools

Adobe: not even once

Very interesting, saving this as I'm planning to migrate to Immich too but I'm still in the uhemm.. conceptual phase...

One question not really related to the core of your post: is running Immich in a VM inside a VM fine in terms of performance? I seem to remember somewhere someone saying that something as resource intensive as Immich would be severely hindered in a VE

The stack is Proxmox running [...] a Debain VM. The Debian VM runs docker where Immich runs.

Running Docker inside a VM is the recommended way to run Docker on Proxmox.

Personally, I just run Docker inside LXC containers. Aparently, this is supposed to be not entirely stable, but its never given me any problems. (I commit an entirely different sin against efficiency, by keeping each service in its own LXC container, each with its own Docker host.)

I also run Immich (in Docker inside a container), and haven't noticed any performance issues. But I'm the only user, and I don't use it heavily, so ymmv.

what zener_diode said.

when i was setting up this version of my homelab (this is v2), i had to contend with how to run docker. The general concensus is that the "best" way to run it is Proxmox>VM>Docker. Proxmox>LXC>Docker is supposed to be viable too. Proxmox>Docker is possible but is not recommend.

I have 20 active docker services with about 30 containers running all sorts of services and they're extremely reliable under this setup. The only time I've had problems is when I ran out of disk space (128gb) because stale docker images weren't pruned.

Yikes.

I’ve dabbled with LR enough to know the basics but haven’t done much in the way of importing files, so most of my photos are sat in storage.

I want to be rid of Adobe entirely.
Is Immich as straightforward to use as LR?

Immich is not like Lightroom at all.

Immich is more akin to Google Photos or Apple Photos. It's a library ONLY. Lightroom is for developing raws, publishing, and storing.

Oh. Ok thanks. I need to do my research.

The open source equivalent to LR would be Darktable.

I see. I'm fairly new to this, obviously. Appreciate the context.

Incredible how software companies can be actively hostile to customers.

what is cc because i dont think you are using c compiler as lightroom alternative

I'm wondering if immich supports reading an existing file structure, so you could maintain the file system and mount it in both immich and next cloud or whatever else?

I think that's possible

CC is "creative cloud" its the Adobe cloud-ified software offering. All the software you used to love, with fewer features and a monthly rate!

Mind sharing which local AI you used, the harness, etc?

Darktable is honestly really good but man is it a daunting learning curve. I've been getting into darktable over the last week and I'm starting to understand the basics haha

Disable every module except like 5 and try to understand those

Hardware: I have a custom setup with a bunch of amd graphics cards. I will do a writeup about that sometime this year.

I use opencode exclusively. It's probably not the best harness but it's the one that gives users the most control and I like that.

The actual models I used for this work are Qwen3.8 27b - FP8 and Qwen3.8 - Flash - Next UD_Q6_K.

I wrote my own agents with duty separation. Flash is the orchestrator and long-horizon guide. While 27b is the faster, more expandable worker. This combination feels about as smart as an opus 4.8 with maybe not as great token efficiency. For simple tasks, i use the worker directly. I have a ton of repo documentation and my agents have access to read and write to Forgejo, both for git and api (issues and PRs respectively)

Maybe I need to write an article about this...

When you say immich has build in tooling for google takeout, do you mean [immich-go] (https://github.com/simulot/immich-go) or is there another, possibly better, way?

You could run it in LXC directly with [community-scripts] (https://community-scripts.org/scripts/immich). I have it running like this for close to a year and am really happy with it. It provides you with a update script that you would use instead of a docker pull

I've been doing this for years too, keeps services isolated and I can share the same GPU with several containers. Works great!

Yes, it does.

I’ve been on a similar journey, but am stuck on a particular issue. We have fifteen or so years of albums in Lightroom that we would use as our source of truth and then export those photos to various photo hosters. I came upon a similar solution as you with extracting things from the catalog SQLite DB and using a naming scheme for dealing with the lack of nested albums. The biggest issue though is the lack of custom sorting. Almost every album uses a custom order to tell some sort of story and doesn’t always follow the capture time of the image. There’s this feature request out there right now, but it doesn’t seem to have much traction. Have you had to deal with or otherwise work around this issue, or are your collections basically chronological?

I haven't had to deal with custom ordering. Everything I've shot can be chronological with the exception of some portfolio albums where I simply forgot order mattered.

I haven't spent a ton of time in Immich's backend but do you know where they would implement such a feature? Would sorting by file name solve this? Does the ui use the api to call pictures in a specific order? It seems like functionality is platform-dependent.

If you wrote an article I'd read it! But this is a great starting point nonetheless, thanks

I used immich-go when I was migrating over, worked a treat.

I use it and it works great for me. My use case is I have several users which backup our mobile phones into Immich. Then I created an additional "user" that is more like a service account. Under this extra account I have it set to scan an external path which I a drive where we upload everything from a DSLR camera. Then I share this users library with my main account so I can browse our DSLR images while on the go through Immich.

no, you're completely correct. I used immich-go which is what I recommend

It was faster to export the 1000 items at a time via the photos web UI which still kept the metadata together.

Apple was not much better, requiring an unsupported third party tool.

I have to give it to them, that truly made me not want to use their products ever again.