TL;DR
- Hubble is a free, open source notes app built on plain markdown and HTML, marketed as “the best notepad for you and your agents”.
- Its design choice is to structure notes for a second reader from the start, so an AI agent gets consistent headers and predictable metadata instead of guessing at intent.
- The cost sits on portability: a format optimised for machine legibility risks depending on the tool that enforces the structure, and stays portable only if the schema stays simple and documented.
- No verdict is reached: the bet suits people with no legacy vault, and the migration story for years of freeform notes is something a launch post cannot answer.
In plain English
Markdown is a plain-text way of marking up documents with symbols for headings, lists and links, readable in any text editor, which is why notes written in it survive whichever app opens them next. An agent is software acting on your behalf while you are not watching, and it has to parse your notes to use them. Hubble’s claim is that if the note format carries a predictable structure, headers in known places and metadata in known fields, the agent can act on a note directly instead of inferring what each block of text was for.
The trade-off is vendor lock-in: the structure only helps if some tool enforces it, and the more elaborate that structure gets, the more your notes depend on that particular product. A schema is the agreed shape of the data, the list of fields and what each one means. Keeping that schema simple and publicly documented is what would let notes move elsewhere later, and that is a discipline a company has to choose to maintain.
“The best notepad for you and your agents.” Hubble is free, open source, and built on plain markdown and HTML, which sounds unremarkable until you notice what the tagline is actually claiming: that a note now has two intended readers by default, you and whatever’s working on your behalf while you’re not looking. Most note-taking tools that added AI features bolted them onto a format designed decades before anyone needed a model to parse it. Hubble is trying to design the format around the second reader from the start.
That’s a real design choice with real trade-offs. A note structured with a parser in mind can hand off cleanly to an agent: consistent headers, predictable metadata, less ambiguity about what a given block of text is for. Freeform notes, the kind most people have been keeping for years in Obsidian or Notion or a plain text file, carry none of that structure. An agent reading them has to guess at intent the same way a new hire would flipping through someone else’s notebook.
The cost is on the other side. The more a format is optimized for machine legibility, the more it risks becoming dependent on the tool that enforces that structure. Plain markdown survives regardless of which app opens it, which is most of the reason it’s lasted as long as it has. A format built for dual readership only stays portable if the schema itself stays simple and well documented, and that’s a discipline a product has to maintain deliberately, not something the format guarantees on its own.
I don’t think there’s a clean winner yet. Hubble is betting that agent-legible structure is worth the coupling for people setting up a note-taking system today, with no legacy vault to worry about. That’s a reasonable bet for a fresh start. Whether it holds for the much larger number of people with years of freeform notes already sitting there depends on how good the migration story ends up being, and that’s not something a launch post can answer. It’s the kind of claim that only gets tested by whether people are still using the format three years from now.