Genuinely curious, maybe I'm missing something. What makes this more "agent-ready" than Obsidian .md files on disk or even something like .org files?
I can let my agent edit those files all the same + there's better integrations and knowledge readily available on these.
I can tell my agent to edit the Obsidian Markdown just like that, but I can also tell it to create a full-on notebook with backlinks and DataView queries. I can tell my agent to create logically complex .org notes with code blocks and references.
To me, Hubble just seems like a Markdown editor, based on the premise that agents are able to edit local files.
Fair comment. Hubble and Obsidian both use Markdown on-disk, so they're equally agent-friendly.
There's a few reasons I built it instead of continuing to use Obsidian (which I've used for years):
- I wanted something closer to Apple Notes or Notion for simpler editing. Hyperlinks are popups instead of in-line strings. Formatting is WYSIWYG, without *_ characters in-line. There's a / menu to discover formatting options. etc.
- I wanted an app that could display HTML pages alongside Markdown. And more interestingly, I wanted to give those HTML pages access to Markdown files in the folder, using Markdown as a mini-database to build apps on top of. This is likely possible with Obsidian plugins (everything is) but I wanted something first-class. The hubble.md homepage tries to communicate this, but I could likely do a better job :)
- More recently, I've wanted markdown -> agent handoff to be better. We've added an embedded terminal to chat with agents alongside notes directly, and we've just added a first-party commenting system similar to commenting in a Google doc (agent accessible with in-file metadata). All of these features have been community-led. I'm very grateful for our repeat contributors.
If any of that sounds interesting, give it a try! Otherwise, feel free to give feedback on the app itself or the homepage on GitHub. All of that is open-source: https://github.com/bholmesdev/hubble.md
Usually, I have vscode/Claude Cowork up on one side and on the other is Obsidian. I made a plugin[1] that converts emails to `.md` so they can also be in Obsidian folder -- so Claude can query those emails the same way it does my notes. Works beautifully.
I feel like there's not enough content on the landing page to help me understand why I need yet another notetaking app. The burden of proof is high given how many of these there are.
Good feedback, I'd like to add more sections to the homepage to describe the components / motivations more clearly. The scope has increased to more agent collaboration features this past month too. Here's the reasons I shared elsewhere here:
- I wanted something closer to Apple Notes or Notion for simpler editing. Hyperlinks are popups instead of in-line strings. Formatting is WYSIWYG, without *_ characters in-line. There's a / menu to discover formatting options. etc.
- I wanted an app that could display HTML pages alongside Markdown. And more interestingly, I wanted to give those HTML pages access to Markdown files in the folder, using Markdown as a mini-database to build apps on top of. This is likely possible with Obsidian plugins (everything is) but I wanted something first-class. The hubble.md homepage tries to communicate this, but I could likely do a better job :)
- More recently, I've wanted markdown -> agent handoff to be better. We've added an embedded terminal to chat with agents alongside notes directly, and we've just added a first-party commenting system similar to commenting in a Google doc (agent accessible with in-file metadata). All of these features have been community-led. I'm very grateful for our repeat contributors.
Why even limit the app to Rust? Cross compilation is for humans. Just have the LLM make multiple versions of the codebase using whatever language and SDK is best integrated into each OS.
Well they can win! Since everyone is so pro LLM don't write apps in Nodejs LOL
Pre LLM nodejs was convenient option to deploy to all platforms at once.
Post LLM that's no longer a limitation! Just prompt more to create apps in native language! Anyways RAM is getting costlier by the day due to LLMs themselves
That is unless devs don't use LLMs at all. Then they should simply state in the GitHub repo that we don't use AI!
I wouldn't use AI for hobby projects tbh. It's my personal preference
> Just prompt more to create apps in native language!
If you have ever done that you know that that's as true as "write once, run everywhere". There will always be tweaks and iterations, even if it saves you a lot of time.
The counter argument is that since LLM resembles magic why use inefficient nodejs that is basically an app running inside chrone
Why not use native binaries?
Now, I'm having fun with tldraw.[1] With their new tldraw offline, I can continue to work in Claude Code (and others) and I can either output or interact/input from tldraw. This is more of a free-form, infinite canvas. I treat it like Excalidraw + Draw.io + Freeform Notetaker.
For text, most agents/LLM/assistants are pretty happy and smart using the same Markdown files to anywhere I have allowed them to interact.
I’ve been waiting for an app like this to be developed! .md files can be opened from Finder! The tree file view is just like I’m used to with Obsidian! Frontmatter support! There’s even an editor toolbar. Almost everything on my wishlist for a Markdown editor, except for Mermaid rendering.
Since this project uses such good agentic engineering practices, I was able to use Claude Code to add support for Mermaid diagram rendering and produce my very own local build on Mac, all in about 30 minutes.
Thank you! Hubble+cmux gets me close to the agentic IDE I’m working towards.
If you don't mind me asking, what are "good agentic engineering practices"? To my knowledge, we are still exploring many techniques of using agents within a codebase, and have not widely agreed on "best practice".
Thanks, glad you like it! Feel free to start an issue and a prototype PR for mermaid diagramming support. It's been on my todos to add: github.com/bholmesdev/hubble.md/issues
But does joblin support anything more than markdown? I'm thinking of using it as a like a store for agent plans, so I can easily edit stuff.
Can an agent add rendered HTML, meramaid etc in joblin?
it's interesting -- i think most (non-technical) people don't really have a muscle memory go-to markdown editor. i certainly didn't. i hated obsidian at first! funny how learning to like markdown makes obsidian like 1000x better haha
The value proposition of Markdown is that you don't need a special editor, you can open it with any text editor that comes with your OS and it's human readable already.
Looks interesting. The dual interface from the start (React UI for humans and just editing .md for the agents with skills) is an interesting idea and probably how a lot of new software projects will work going forward
I feel like tables are generally underrated while the AI ecosystem is focused on .md files but I guess markdown has tables so they’re included to some extent
I think that is a really nice implementation. A few questions:
Can you easily share lists or pages or is it all local? (Is there a cloud element to it)
What are the most useful examples you’ve personally found when it comes to using the HYML component?
I built https://smalldocs.org for my agent’s Markdown needs. I extended it so it can render complex formats: Mermaid diagrams, charts, in-document slides and even in-document spreadsheets. It gives your agent a good selection of formats to express itself as it helps you understand whatever topic you’re looking into.
It has a rest api and a cli tool, and also the possibility to make headless deployments, and it can sync clips without a sync server. It can serve folders/tags and optionally expose a read or read/write api for JSON files on the same level. It supports lua plugins, and I have a fal.ai plugin to generate/edit images already in the repo (some other examples there).
I think it looks high quality. Perhaps it needs a way for those mini-html apps to store data? Also some additional docs and clear marketing that differentiates it from Obsidian would be cool as well.
Thanks for sharing this handfuloflight! This is Ben, creator of Hubble. I've been replying to comments throughout this post (appreciate the thorough feedback). If you have any questions for me, fire away.
Cool, I needed a lightweight markdown editor.
I like Obsidian but it's not really a notebook.
What's missing is a ToC view for your .md as a sidebar which Obsidian has.
There's even a community plugin MCP that makes this easier. Afaik the hot reload is the only thing that "beats" this but I haven't had issues with Agents interacting with obsidian through the MCP - I'm not sure how much more useful it is to talk with an agent through a note taking app vs through its CLI.
Obsidian does hot reloading too. Usually, I have vscode/Claude Cowork up on one side and on the other is Obsidian. I made a plugin[1] that converts emails to `.md` so they can also be in Obsidian folder -- so Claude can query those emails the same way it does my notes. Works beautifully.
You could at least read the tagline or something, before leaving low-effort comments. Im a big fan of Obsidian, use it every day, but if Im looking for something "backed by Markdown and HTML" then obviously its not even an alternative...
The shared markdown files between human and agent notes is a nice touch — no proprietary format lock-in, and agents can just read the files directly. Does it handle conflicts when both the user and an agent edit the same note at the same time?
I can let my agent edit those files all the same + there's better integrations and knowledge readily available on these.
I can tell my agent to edit the Obsidian Markdown just like that, but I can also tell it to create a full-on notebook with backlinks and DataView queries. I can tell my agent to create logically complex .org notes with code blocks and references.
To me, Hubble just seems like a Markdown editor, based on the premise that agents are able to edit local files.