Show HN: Nodini.ai – AI-Powered Brainstorming with Branching Conversations
nodini.aiShow HN: Nodini.ai – AI-Powered Brainstorming with Branching Conversations
Hi HN! I’m the maker of Nodini.ai (https://nodini.ai/), a tool I built out of frustration with how linear most AI chatbots are. I found myself constantly losing track of side questions or ideas. So, I created a system where you can:
- Start a regular Q&A with the AI - Branch off any point in the conversation to explore tangential ideas - Visualize everything as a node-based graph
What’s Different?
- Branching Discussions: You’re not stuck in a single chat thread. Each conversation node can spawn multiple sub-branches. - References to Previous Nodes: Type @nodeId to reference earlier parts of the conversation so the AI can consider them. - Visual UI: See your conversation flow at a glance and jump between branches easily.
Why I Built It
As a developer, I often juggle different directions or possible solutions. Traditional AI chat sessions felt too linear—once I pivoted to a new angle, I’d lose context from previous ideas. I wanted a tool that keeps track of every thought process in one place, so I made Nodini.ai.
Try It Out
- Live Demo here: https://nodini.ai/ (No login required; you can use your own OpenAI API key to power the AI. The key is stored locally and never sent anywhere except OpenAI’s servers.) - If you’d rather not use a key, there’s a read-only demo mode showing how branching works.
Tech Details
- Built with React + TypeScript. - Uses a simple graph structure to store each Q&A node. - References are handled by passing the content of linked nodes back into the prompt. - All chat graph data is stored in the browser’s local storage, meaning conversations persist on your device but are not synced to any cloud or external database.
Notes
- Currently optimized for desktop—the mobile experience is still a work in progress and may not be fully functional. - No backend storage—your conversation history stays local, so clearing your browser storage will erase your past chats.
I’d Love Feedback On
- The overall concept—does branching help your workflow? - UI/UX design and how we could make the node visualization better. - Any additional features that might help (e.g., advanced search, embedding code snippets, etc.).
Thanks for checking it out, and I’m happy to answer questions or hear suggestions!
(No friends or fans were harmed in the making of this thread—i.e., I haven’t invited any outside booster comments. I look forward to honest HN feedback.)
In the visualization on your site, I see some curvy arrows linking between two different branches of the conversation. Does that mean anything in the context of the chat with the LLM? Or is it like some detected similarity between the two nodes that was determined after the chat occurred?
I've been dreaming of this in context of an established lesson as the root