Are there any other similar, long-term partnerships that have not resulted in a merger?
Yes, e.g. Nikon (Canon, Sony, …) and Adobe 
I’ve got affinity 2. Didn’t take affinity canva.
Export to application still show affinity2, but export to affinity does not.
Does this mean I have to create a Canva account, or should Affinity 2 work ?
First choice could explain the deal between DxO and Canva (canva side).
EDIT : Seems affinity canva is mandatory. Extensions are not the same .afphoto vs .af, it’s a genuinely new format with a different internal organization.
I’m a bit worried about this Canva account. Software can’t be free; there must be a catch.
Here are informations I found, they might ask for verification.
Affinity 2 vs canva :
1. Canva’s Privacy Policy (very different from the old Serif one)
| Old Affinity (Serif) | Affinity by Canva | |
|---|---|---|
| Data storage | UK/EEA only | US, Australia, Singapore, EU, UK, Philippines… “and any other country” |
| Third-party sharing | No | Yes — advertising partners (Facebook, Google, Samsung TV, etc.) |
| Audit | No | Canva can send an agent to audit your computers (clause in ToS) |
| Data deletion | Simple | Limited — no control over “global” transfer |
2. No perpetual license
With the old Affinity, you owned the software. Now Canva can modify, restrict, or revoke access at any time. It’s a service, not a purchase.
3. AI / model training
Canva states it doesn’t use your Affinity files to train its AI. But:
-
The default account settings are opt-in (not opt-out) for AI training on Canva content.
-
One user reported being re-opted in even after disabling the options in the app — they had to change the settings in their Canva online profile.
-
The legal documents don’t explicitly restate the “we don’t train on your Affinity files” promise — it’s a public statement, not contractual.
4. Usage data collection
At installation, a checkbox “I accept to share my usage data” is checked by default. You need to uncheck it manually.
In summary
The trade-off isn’t “we sell your photos” — it’s rather:
-
You accept Canva’s ToS and privacy policy (much broader than the old Serif’s)
-
You lose the owner status in favor of revocable access
-
You need to check and disable several default options (AI, usage data)
-
You implicitly accept that Canva can change the terms in the future
Same here. Since Affinity Photo 1. My only use has been for sky replacement. I was taking night sky photos on a tracker - which would smear my foreground a bit, so I’d take two photos, one with tracker on (sky), one with tracker off (foreground) and replace the star-trails sky with the on-tracker one. I mostly don’t bother these days (faster glass and better lowlight camera).
As to the mechanics, I’d export TIFFs, open them in Affinity. I don’t really see how exporting an .AF file changes that. Maybe color space - although even AP1 was fine with TIFFs with color space information (and I wasn’t up against the boundaries of even sRGB) so…(?)
Whether it’s Affinity, Capture One, or Nik, I do PL export in 16-bit TIFF using suitable colorspace (mostly sRGB), exit PL, and only then edit the output.
An old way of doing things, but it keeps me away from OS resource problems…
Another good point is that in this way you don’t have to mentally switch between the software too often.
I don’t think I would ever need masks created by PL in other programs for special corrections, not available in PL. Perhaps it would be useful in Affinity for layer merging, but that’s too complicated for me to do. If the image requires too complicated editing, I delete it – it simply means for me that the input is bad.
Bottom line: try NOT to use two resource hungry apps in parallel. Keep focused.
I do this too, till now.
Depends of the kind and usage of the image.
I most often make photo for my pleasure (generally, not only, worldwilde photography).
But I use them for my job too (Creating HDR environments for 3d lighting in sfx, creating base for texturing, and sometime products. Those 2 last kind benefit from masks for changing colors, hue, etc … long after PL job is finished in 2d or 3d softwares. Here embedded masks are more than welcome).
For what purpose that you can’t do in PL?
Just trying to learn something new…
I understand why the rest of Canva’s policies might be concerning, but usage data? I used to write software - we used aggregated usage data to figure out what parts of a particular app were most used (AKA ‘hotspots’) and focus resources on making those parts faster - most used hotspot at the top of the list.
Maybe I’m being naïve here, but can anyone point out how usage data could be used for nefarious ends?
As a simple for instance: Compositing. If I want to cut out a part of an image to place it into a new background, the PL mask allows me to do that ease. And I cannot do compositing within PhotoLab.
There are many other examples, but that one seems the most obvious.
OK, and it’s obvious indeed.
Because in the cases I talked about, the workflow (team workflow) does not involve Photolab nor any raw editor. But uses RGB images of any kind.
I was just putting here a summary that I thought exaustive. And indeed some of the bolded text is poorly chosen, and this point should not be emphasized. I’ve just corrected it. Not bold now.
While Affinity Photo 2 still works, it’s now a legacy application, i.e. Canva are never going to develop anything new for it. Similarly, since it’s a legacy application, no other company is likely to develop new features in their software that take account of AFP2. Given that, the chance of this feature ever working for AFP2 is close to zero.
the chance of this feature ever working for AFP2 is zero.
I hope Affinity will donate their in painting brush and clone tool to PL, they are SO MUCH better to use - the PL ones are not very good at all In my opinion. Would be nice to export to AF and then export back again after you have used the in painting brush or clone tool.
I think exporting makes sense if you can’t offer all the features yourself. But then it should be in a format that various photo editors can understand, similar to the ODF format for word processing. Only then can the independence of each tool truly be guaranteed in the long term. It’s also conceivable that the ODF format could be extended for photo editors.
Currently, every tool involved in the workflow must always support proprietary formats.