Ask HN: How do you integrate AI assistants into your note taking?
I've been rethinking my personal knowledge management setup and realized my workflow is becoming AI-first. Most of my projects and ideas now start with a conversation in Claude/ChatGPT, but then I switch to Obsidian for structured notes. This creates friction - my thinking process is split across tools.
Do you "Add AI to a note taking app" or "Add structure to AI apps"?
I'm curious if others are thinking about this and how others are handling this.
I've been experimenting with an evolving framework called think-center[0] that directly addresses this friction. It treats AI conversations as structured 'council sessions' with defined perspectives, and includes Obsidian integration - basically your vault becomes the shared workspace.
The key insight: instead of 'AI + notes' or 'notes + AI', it's 'AI as part of your PKM system'. Sessions auto-save to your vault, perspectives can reference your existing notes, and everything stays in sync.
What makes this actually work: treating the perspectives (Weaver/Maker/Checker) like intelligent collaborators on your team, not command interfaces. When Checker raises objections, that's the system working - debate which ones matter. This collaborative approach transforms output quality.
The friction points reveal where the magic happens. If a perspective annoys you (usually Checker), that's your cue to engage deeper with its purpose rather than bypass it.
Still evolving based on what works, but it's been solving the exact split-brain problem you're describing.
[0] https://github.com/achamian/think-center
I've been building real-time collaboration into Obsidian (plugin is called Relay [0]) with the intent of integrating AI via standard collaboration flows.
We haven't released our AI stuff yet, but have done several experiments and it is looking really promising.
I'm also a heavy Zed user. A bunch of the UI (pre-Agent-overhaul) has been inspiring our work.
I've also experimented with voice recordings when on the go, and low friction transcription has felt like one of the killer apps.
One cool thing our plugin supports out of the box is using the voice input on Android for real time text-to-speech on Desktop (just have the note open on both devices).
I've also enjoyed using a script [1] I wrote to summarize bash history each day into my daily note.
It feels like there is a real spark in the Obsidian ecosystem right now. Super fun time to be building!
[0] https://relay.md
[1] https://notes.danielgk.com/Software/obsidian-import
Ooh. I love the voice recordings to solve the mobile/on-the-go flow. In practice, it is often indeed small things.
Interesting idea of adding AI via a real-time collaboration plugin. I'll try to keep an eye on it :-)
I'm playing with the idea of using a code editor like Zed.dev for my notes, and using its "AI Agent" flow.
I'd move the Obsidian markdown notes into a git repo (for versioning). Which would also open the door to running Claude Code to analyse my note taking.
Creating commits could be (semi) automated with a script that takes a git-diff and ask a local (or cheap) LLM to create logical commits with a reasonable commit message. The script could run on file save or periodically.
Mobile editing would be worse than Obsidian's iOS app. Workingcopy.app could provide an inferior but working UX.
Thinking in progress :D