Bear CLI is fantastic news! This unlocks a bunch of workflows I’ve wanted for a while.
One small “might be useful to someone” note: a while back I built a hobby project called mde, a terminal Markdown preview/editor. It can pair nicely with bearcli if you ever want to quickly view or lightly edit note content without leaving the terminal (or prepare Markdown to send into Bear).
It was a little tricky to get Claude Connector setup, because the “save” button was greyed out with the suggested path to bearcli. But once I added a space and then backspace, the save button was available and everything connected fine.
I’m really liking this. I’m not a super-techy guy but it’s simply to get working.
I just extended my Bear subscription for another year.
@matteo any chance that you add a backup function to bearcli to create a backup with it?
Besides that I would like to know if there are any plans to add an automatic backup feature for the Bear apps on iOS and MacOS. To think about manually backing up my precious data drives me crazy.
Probably not. Bear is local first and also private-oriented. Safely exposing your notes over HTTP is not trivial. We would have to add authentication, and if you want to access it on ChatGPT, you would also need to expose it to the world through a tunnel like NGrok. That said, we follow how the field develops, and there might be other solutions coming up.
We have considered it, but it is technically a bit complicated, so we are starting without. We have also considered adding export (to different formats), and syncing with iCloud. This is a first version, and we want to learn more about how it will be used.
My request is for localhost-only HTTP (like 127.0.0.1:6190), not remote access. According to the chat app I use (FlowDown) iOS/macOS app sandboxing blocks stdio, so HTTP is the only viable transport for sandboxed MCP clients like FlowDown.
Would you consider a small helper app that bridges stdio to HTTP locally?
@matteo I was asking about the x-callback API you have – wondering if you have any plans to update it. But I guess now you have more Technically, the CLI becomes an API too.
This returns each note as a separate JSON object, one after another. Since the output isn’t wrapped in a JSON array, it’s not valid JSON as a whole, which makes it awkward to process.
Would it make more sense to return the results as a single JSON array of notes instead? That way the output would be valid, parseable JSON out of the box.
Thanks for the response. Personally I would prefer if the native app could handle that. Having a consistent and permanent backup of the data outside of Bear on iOS and Mac is a feature that I deem as necessary before putting a lot of valuable data inside Bear.
Bear.App should have a setting and should manage the Backup automatically. Currently I wouldn’t know where to look for my Bear data if something goes wrong. Unregularly, I think about backups and do a manual backup. But this isn’t ideal.
Technically, the CLI becomes an API too.