About Lattice

One person, one app, and a list of the things I am not going to tell you.

Last checked 9 September 2026

Who

Jarrel Pothoff, in the Netherlands. There is no team, no company, no investor and no roadmap meeting. When this site says "I", that is accurate; when a one-person project says "we", that is the first thing you should catch it on.

Why it exists

Two things I wanted were never in the same app: notes in a format I could take anywhere, and links between them that were actually load-bearing rather than decorative. Markdown tools treated linking as a plugin; linked tools treated the file format as an implementation detail you were not supposed to ask about.

Lattice starts from the link. The graph, the backlinks and the search all read the same [[wikilinks]] you write while thinking about something else — so there is no separate filing step, because a structure you have to maintain is a structure you eventually abandon.

Choices I made on purpose

  • Markdown content, not a proprietary document model. The exit has to work, and a format every other tool already reads is the only exit that does.
  • Encryption with names you can look up. AES-GCM-256, PBKDF2-SHA256, keys in the system Keychain — three terms you can search, rather than a padlock icon.
  • No analytics. Not "privacy-respecting analytics". There is no analytics SDK in the app at all, which is a thing you can check by grepping the source rather than a promise you have to accept.
  • Two platforms done properly over four done thinly. Mac and iPhone share one Swift core. The browser build in the repository is a different codebase, and I count it separately.

What I deliberately do not claim

This is the section that matters, and it is here rather than in small print because you will find all five out anyway — better from me than from a disappointment after installing.

  1. Lattice is not in the App Store. There is no build you can install, no TestFlight, and no date. The waitlist is a mailing list, not a queue.
  2. Your notes are not loose files in a folder you control. They live in the app's own store; the content is markdown and you can import a folder of .md and export back to one. If files-on-disk is your requirement, that is a real difference and you should weigh it.
  3. The AI layer does not work yet. The vault-chat endpoint and the "related notes" call exist in code, the embedding service is close to a stub, and the backend that would answer them runs nowhere public. In a build today those features fail quietly.
  4. The platforms are not feature-equivalent. Live collaboration and the sandboxed plugin API exist only in the browser build. The graph and the encryption live in the native core. Saying "everywhere, identical" would be measurably untrue.
  5. There is no company. No Chamber of Commerce registration, no VAT number, no registered address to publish — which is also why the machine-readable data on this site describes a person and a brand, and not an organisation.

Everything above is checkable in the source, and the trust page says how — including the parts about this website rather than the app.