Hey Product Hunt 👋 I built SlateKore out of frustration. Over time, my research workflow became deeply AI-driven. I was reading dozens of papers, exploring models, running experiments, and using LLMs daily to summarize and synthesize ideas. I use Obsidian as my knowledge base and when I added Gemini CLI into the mix, things became incredibly powerful ⚡ But also… chaotic. Gemini CLI made it possible to: * Fetch and summarize papers instantly 📄 * Generate structured outputs * Refactor notes * Synthesize across projects And Obsidian gave me: * Local-first ownership 🔐 * Deep linking 🔗 * Graph-based thinking 🕸️ * Markdown native flexibility ✍️ Together, they’re insanely powerful 🚀 But here’s the problem: AI amplifies whatever structure you have. If your vault is messy, AI just creates structured chaos faster. My notes started duplicating. Context windows exploded 🧠💥 Summaries weren’t connected to projects. Outputs lived in random folders. I was moving fast but not compounding knowledge. So I stepped back and rebuilt my vault from scratch with one principle: **Design it for agents first, humans second.** SlateKore is that foundation 🧩 It’s an opinionated research operating system for Obsidian built specifically so tools like Gemini CLI can operate cleanly inside your vault. Instead of random prompts and scattered outputs, you get: * Clear project scoping 🎯 * Typed research objects (papers, models, datasets, codebases) * Structured synthesis layers 🏗️ * A predictable capture → analyze → build loop 🔁 * AI outputs written back into the right place 📂 Right now, SlateKore focuses on structure and workflow discipline. It doesn’t try to be magic or fully autonomous it makes your knowledge system agent ready 🤖 Because once you give AI even a little autonomy fetch sources, summarize, link notes structure becomes everything. It’s fully open source ❤️ I believe: * Researchers and builders should shape their tools * AI-native workflows shouldn’t depend on SaaS lock-in * The future of research is local-first + agent-augmented If you're using Obsidian and experimenting with Gemini CLI (or any LLM CLI), I’d genuinely love your feedback 🙏 What breaks in your workflow today? What feels messy once AI enters the picture? Thanks for checking it out 🚀

Slatekore

AI Research OS for Obsidian Power Users

AI Research OS for Obsidian Power Users