My Thoughts on the New License Model (Customers now pay $49 annually on top of the purchase price just to receive updates)
I’ve been a loyal DEVONthink customer for over 10-15 years, purchasing every major version upgrade. I’d like to share my perspective on the new annual licensing model in version 4.
My Main Concern: The License Model
In my view, this is effectively a subscription model - even if not called that way. While we can continue using the purchased version, we simply don’t get any more updates without paying $49 annually. To me, that’s a subscription model in different packaging.
I liked the traditional model much better:
One purchase = all updates until the next major version
Major version upgrades (v4 → v5) are paid, which I think is completely fair
Users get updates, bug fixes, and compatibility patches for the entire version cycle (which has been around 6 years between major versions)
I appreciated paying once and knowing what I was getting.
In my view, the annual payment model doesn’t seem to fit a local application with no ongoing service costs. I understand you call it “modern” and “flexible,” but it feels like the main motivation might be revenue-driven rather than user-focused. This doesn’t fit what I thought DEVONthink’s company philosophy would be.
Apart from that - Version 4’s Features
The AI integration, which seems to be the main selling point, doesn’t appear to deliver what I would consider the next evolutionary step:
RAG (Retrieval-Augmented Generation)
Vector embeddings or semantic search across my entire database
Meaningful context window for analyzing large documents
Correct me if I’m wrong:
The “natural language search” just converts your query into Boolean syntax for the old search engine - not true semantic search
When documents are too long, the app tells you to use the regular search instead - which defeats the purpose of AI integration
The built-in machine learning search engine (the “See Also” feature, document classification) remains unchanged from version 3 - no improvements to the core AI (machine learning)
It looks to me that the main novelty is not much more than an API wrapper to external services like ChatGPT/Claude. It’s called DevonThink - From what I read about new features, I don’t see much more thinking happening than in older versions.
I personally don’t see any substantial innovation that justifies both the upgrade cost and the new annual payment model.
We’ve been over that ad nauseam. It is not “annual”, but “for one year”. If you pay every year or every second, third or tenth year doesn’t matter. When you pay, you get the then current version including all updates it had received since you last paid. And you receive all updates in the next 12 month. If you don’t want to pay, then don’t and use the current version.
Except that it is the company that defines what a “major update” is. If they behave like Apple or Adobe, you get a major upgrade every year. Which difference would that make to the current model?
You are certainly welcome to your opinions and whether you’ll decide to upgrade. That being said, time marches on and version 3 will not be developed over the long-term.
By the way, you are coming at the topic with your own preconceived notions, an approach that often leads to disappointments when things don’t appear as you imagine they will. Also, you are incorrect in guessing the internal AI has no improvements. 23+ years later it continues to be refined. External AI integration is discussed in detail throughout the manual, starting with the Getting Started > AI Explained section. Lastly, version 4 is nascent, not even being public for a year yet. There are things still to come.
Correct on the technical definition. But let’s be honest: ‘flexible’ doesn’t mean much when the practical choice is either pay or get left behind. And yes, major versions are company-defined - which is exactly why I preferred the old model where they had a 6-year track record.
I appreciate the response, though ‘preconceived notions’ feels a bit dismissive of specific technical feedback. If there are concrete ML improvements beyond v3, I’d genuinely like to know what they are. And ‘things still to come’ is fair, but I can only evaluate what exists now.
This is a business, i.e., a for-profit entity, providing for the employees and their families for 23+ years now. This is our livelihood and the fruit of thousands of hours of work and testing and fixing and retesting, etc.
I’m not questioning your right to profit as a business.The fact that DEVONthink never adopted a subscription model - even when subscriptions became common in the industry - led me to believe this was part of your company philosophy. I may have been mistaken about that.
This is the feedback section, so I wanted to share my perspective. What you do with it is up to you.
And your comments are fine to speak aloud here. It’s a free place to discuss.
And whether you feel it’s merely semantics, we don’t. As mentioned often, we don’t have a subscription. A subscription terminates access to something. For example, Sirius XM keeps calling me because they want me to renew a subscription. I won’t and I will no longer be able to use their service. If they said, “Well, you can keep listening to the stations you already have, that’s fine but if we add new features and stations, you don’t get to have them too!” then that would be more akin to our model. (PS: This approach is not our invention so we are certainly not the only ones using it.)
I have never seen any post in any forum for any product change the owner’s mind about price. Of course, a bit of grumbling and harumphing and how-dare-ya is always interesting and diverting. But, pay to play is the rule, as it should be. We all do our own calculus.
@JohannesD Just look at it this way: life has become more expensive for everyone in recent years. That includes the people at DT. However you feel about the new features in DT 4 and whatever you think about “modern” and “flexible” regarding the new pricing model, the fact is that the DT team wants and needs to generate more revenue. That’s their right.
Even if not everything is perfect, DT is the best app of its kind. That hasn’t changed with version 4.
You and every customer have two options.
You look for another app. Fortunately, DT makes it easy to export all your data.
You accept that DT has become more expensive, but are happy that the best app of its kind is constantly being developed and that you can use it. Hopefully for the rest of your life.
