undefined

points

[-]

It would be nice if you could “see” the AI in your vault making changes. Almost like a Google doc collab session. Even if you weren’t directly interacting with the agent, and it was making change thru a CLI/MCP, its presence would be highlighted in the frontend. And then it appears as its own contributor in the git history.

by tarr11242 hours ago|

parent|

[-]

Git is great as the durable layer, but what fiatpandas is pointing at is a bit different — real-time awareness that the agent is active right now, not a retrospective diff.

I've been working on this split in a side project (https://github.com/rillmd/rill — vault layer on top of Claude Code). Git still handles the durable side for free, since the CLI agent just writes files and commits normally. The live side comes from Claude Code hooks (UserPromptSubmit / Stop / PostToolUse) appending to a plain activity-log.md that the Electron GUI tails. Cruder than Google Docs cursor presence, but cheap — and the log itself is just another markdown file in the vault, same data model as everything else.

One trade-off: going the other direction and letting the frontend detect edits via FSEvents-style watching runs straight into the reconciliation issues the Zettlr subthread is about. Hooks on the agent side are less elegant but sidestep the whole class.

by crashabr1 hours ago|

parent|

prev|

[-]

I'm following a startup who does just that: https://www.moment.dev/docs

It's still early stage, but I love their pitch so I'm following them with fingers crossed.

by lucaronin8 hours ago|

parent|

prev|

[-]

the idea of being git-first it's exactly for that! so you can setup AI as a git contributor and see its changes clearly