You might assume that apps will do certain things automatically. However, many decisions are yours to make, not the app’s. We prefer clear opt‑in controls rather than hidden opt‑outs. Here are a few optional things that are up to your control in DEVONthink or DEVONthink To Go. (more)
Using the Model Context Protocol (MCP) you can make the data stored in DEVONthink and many of its abilities availble to your preferred AI client, e.g., Claude or Codex. But how does MCP work? We have made two new tutorial about it for you. (more)
MCP (Model Context Protocol) is a mechanism that allows AI assistants to communicate with apps to accomplish tasks for you. On macOS Sequoia and later, DEVONthink now has its own MCP support built-in that lets your preferred AI agent access and work with your databases. (more)
AI inside DEVONthink can be a powerful tool. But when it comes to AI, some users have security concerns, such as possible prompt injections. So, what exactly are they, and are they a risk in DEVONthink? (more)
If you want to take notes about a document in your DEVONthink database, there is a feature for exactly that: the annotation. This is a specialized note document attached to another document. Here is how to create and use them. (more)
The popular AI app Ollama, typically used to run AI models locally on your Mac, now offers cloud models. This gives you remote access to many models, even if your hardware is limited. And you can also use these models in DEVONthink and DEVONthink To Go. Here’s how to set them up. (more)
Access to external AI is an optional part of DEVONthink and now DEVONthink To Go. But some people have wondered, “Well, what would I actually use it for?” On the Mac Power Users forum, a user posted her real-world use case, which is an excellent example for AI use. (more)
If you’ve started using generative AI features in DEVONthink, you’ve probably noticed that the results vary. Crafting a good prompt can make the difference between a helpful and a vague or off-topic response. Here’s how you can create effective, but still reasonable, AI prompts. (more)
If you want to use AI in DEVONthink, you can choose between server-based and local models. Whether it’s for privacy reasons or just the novelty, running AI locally on your Mac may be something you’re considering. But how does that work and what options are available? (more)
In many business areas, important knowledge — such as programming code or strategic considerations — is generated in AI chats today. But how safe is that knowledge there? Marcus Schuler addresses this question in the German-language newspaper FAZ and also highlights DEVONthink as a safe archive. (more)