Kepler vs. Obsidian: the note vault you maintain vs. the memory that maintains itself
Obsidian is a powerful local note vault you build and tend by hand. Kepler is lightweight, URL-first memory that captures what you find and turns it into AI-usable context — with no vault to maintain. Here's how they compare, and when each one fits.
Obsidian and Kepler both promise to help you keep what you find and think — but they’re built on opposite philosophies, and which one fits comes down to a single honest question: do you actually enjoy tending a system, or do you just want the payoff without the upkeep?
A useful way to frame it: Obsidian is a workshop you build by hand — a local vault of notes you write, link, and maintain into a “second brain.” Kepler is memory that builds itself — you save what you come across as you browse, and it becomes AI-usable context automatically, with no vault to maintain. Both are good; they’re just built for different people and different jobs.
What Obsidian is great at
Obsidian is a genuinely excellent tool, and for the right person it’s hard to beat.
- You own everything. Your notes are plain Markdown files on your own machine — local-first, offline, and yours for the long haul. No lock-in, no cloud dependency.
- Linked thinking. Bidirectional links, backlinks, and the graph view let you connect ideas over time and see the structure of your knowledge. For people who think by writing and connecting, it’s powerful.
- Endlessly extensible. A huge community-plugin ecosystem and features like Canvas mean you can shape Obsidian into almost anything.
The model is author and interlink: you write notes, connect them, and cultivate a vault that reflects how you think. For writers, researchers, and note-taking enthusiasts who love that craft, Obsidian is a joy.
But that same model is the catch. A vault is only as useful as it is maintained — the links, tags, folders, and notes are all work you do. It’s not built to capture the web as you browse, and it isn’t designed to feed your AI tools; you can bolt AI on through plugins, but that’s a DIY project, not the point of the product. Plenty of people set up a beautiful vault, then quietly stop tending it.
What Kepler does differently
Kepler starts from a different premise: most of what you want to remember already lives at a URL — an article, a post, a paper, a YouTube video, a doc — and the goal isn’t to hand-build a knowledge system, it’s to use that material with the AI tools you already have.
Three things make it different:
- Save where you browse — then forget it. Capture is a tap from the web extension or your phone’s share sheet, not a writing session. Kepler pulls the usable context out of each save automatically (title, summary, text; a deep extract for heavy sources like long PDFs or transcript-less videos). Your memory builds in the background.
- No system to maintain — and nothing new to write. There are no Markdown files to learn, no folders to curate, no links to wire up, no structure to set up. And you’re not building another source of truth: if you just want to save links as context, that’s exactly what Kepler is for. Kepler auto-organises what you save and retrieves what’s relevant for each question — so there’s no cap on how much you capture, and no maintenance tax for capturing it.
- Built to feed your AI. Your saved context isn’t trapped in one app. Through MCP, the same Kepler space plugs straight into ChatGPT, Claude, and other assistants — so you ask questions grounded in what you saved, in whatever tool you already use.
Where Obsidian asks you to build the knowledge system, Kepler is the system — capture, organise, and retrieve happen for you.
Side by side
| Obsidian | Kepler | |
|---|---|---|
| Core job | Author & interlink your own notes | Capture web sources as AI-usable memory |
| Mental model | A workshop you build by hand | Memory that builds itself |
| Capture | Write notes (or clip pages to Markdown) | Save any link in a tap — auto-extracted |
| Organisation | Manual: folders, tags, backlinks you maintain | Automatic: Kepler organises + retrieves per query |
| Maintenance | You tend the vault | Save and forget |
| Storage | Local Markdown files you own | Cloud memory, synced across your devices |
| AI use | DIY via community plugins; not the core purpose | Native: feeds ChatGPT/Claude via MCP |
| Best for | Writers & thinkers who love a hand-built system | Save and forget — capturing what you find to ground your AI |
So which should you use?
If you love the craft — writing notes, linking ideas, owning local files, and shaping a system that’s entirely yours — Obsidian is superb, and nothing here is a knock on it. If your useful material arrives as you browse and you’d rather it just be there, organised and ready for your AI without any upkeep, Kepler is built for exactly that.
And for a very common type of person — the one with a half-finished vault they meant to keep up — Kepler is the honest replacement: the payoff of a second brain without the second job of maintaining one.
They can also coexist. Keep Obsidian for the deliberate writing and thinking you do by hand; let Kepler quietly capture the web sources behind it and make them usable across every AI tool you touch.
Kepler is complementary to your AI tools — it gives them better context from the links you save. It isn’t an LLM, and it’s more than a bookmark manager: the point is using the information inside what you save, not just finding it again.
