Does the AI run on my device or on a server?

Short answer

Most AI journals process your entries on a server. A few run smaller models on your device. On-device is the most private option, server-side is usually more capable, and it is the first thing I would ask any app: where does the AI actually read my entries?

In depth

On-device AI keeps your words on your phone. Nothing leaves unencrypted. It is the most private option, but the models that fit on a phone are smaller, so the insights are usually shallower. Server-side AI sends your entries to a server to be read and analyzed. It is more capable, which is why most apps with real emotion analysis and chat do it this way. The cost is that, for that moment, the company can read your text.

There is no honest way around that second point. If an app does deep analysis of your writing on a server, your entries are readable there while it works. What separates a trustworthy app from a careless one is what happens after: whether it sells your data, trains on it, or shares it.

Pensio is server-side, because the emotion extraction and the advisors need to read your writing to be useful. I would rather say that plainly than dress it up. The promises that are in our control: no selling your data, no training without consent, encryption at rest and in transit, and export and delete whenever you want. For the genuinely local-first crowd, the Obsidian plugin keeps your plain-text files on your own machine.

Ready to try it yourself?

No credit card required. Setup in 2 minutes.