Spine AI – Visual workspace to think across multiple AI models
Hey HN – Akshay & Ashwin here, founders of Spine AI (YC S23). We’d like to show you our new product: Spine Canvas. TL;DR: Spine AI is an infinite visual workspace where you can collaborate with 300+ AI models and agents to think and create better than you would in a single chat. You can try it without signing up: https://app.getspine.ai/guest Backstory: We started building multilingual RAG in 2022, got into YC S23, and through multiple iterations learned that users want iteration and control - not just another chat interface. ChatGPT was originally meant to collect training…
In plain words
Spine AI is a visual workspace that lets users work with over 300 AI models and agents simultaneously on a single canvas. Unlike traditional chat interfaces, it allows users to iterate on ideas, maintain control over context, and collaborate across multiple AI sources in a non-linear environment. The platform is designed for anyone who wants more flexibility and creative control when working with AI tools beyond simple question-and-answer interactions.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
Hey HN – Akshay & Ashwin here, founders of Spine AI (YC S23). We’d like to show you our new product: Spine Canvas. TL;DR: Spine AI is an infinite visual workspace where you can collaborate with 300+ AI models and agents to think and create better than you would in a single chat. You can try it without signing up: https://app.getspine.ai/guest Backstory: We started building multilingual RAG in 2022, got into YC S23, and through multiple iterations learned that users want iteration and control - not just another chat interface. ChatGPT was originally meant to collect training data for GPT’s next version — but it blew up because it made AI accessible. While chat is great for Q&A, it’s terrible for thinking: you get stuck in linear threads, the only way to iterate is to send another message, and you can’t control how much of the previous context the model actually remembers. Each model also has its own “superpower”: Claude is great for coding, Gemini for critique, Grok for marketing. I often use different LLMs as advisors, because each brings a slightly different lens – useful when you’re thinking deeply, exploring tradeoffs, or checking blind spots. So we built Spine: Spine Canvas lets you think in blocks instead of threads. You can connect blocks, which ensures implicit context passing. You can also explicitly refer to connected blocks as part of the current block’s parameters. We have dedicated blocks for: Image generation, deep research reports, slides, chats, prompts. All of these can pass context seamlessly between each other. Because it’s a canvas, you can branch — explore rabbit holes without losing the main path. You can swap models freely. Lists are first-class citizens: run the same process across multiple items (e.g., “deep research these 10 companies”) simultaneously. You can change a single block and trickle updates through the entire branch. To help with the blank-canvas problem, we have an AI assistant that helps you get thinking right away by building you your initial canvas. You can also bring in context from outside – chats from other LLMs, YouTube videos (with transcripts), or web pages – and treat them like native blocks on the board. We also have Miro-style notes and shapes to let you add human context directly into your AI workspace. How we use it (and who else does): The simplest way to use Spine is to chat with the latest models, branch the conversation at any point, and continue it with different models on the same board. As a visual founder, I use Canvas to map GTM, product priorities, and engineering decisions. For example, the prompt engineering for our upcoming App block came from exploring how six different LLMs reasoned about the system prompts of other AI tools – all in one board, in under 20 minutes. I hit “flow” a lot more because everything lives in a single workspace. Founders have been surprisingly early adopters – probably for chaos-meets-structure reasons. Product folks, especially those who use tools like Miro or FigJam, use it for strategic planning, prioritization, and finding blind spots or connections in their roadmaps. Investors use it for market research since our deep research reports are more comprehensive and easier to iterate on than the chat LLMs. And we’ve seen some power users use it for learning new concepts. The Ask: Give it a try and tell us what you think. You can use it right now without signing up: https://app.getspine.ai/guest If you create an account, there’s a free plan with 2,000 monthly credits + 500 daily (credits scale with underlying model costs). If you want to chat: [email protected], Discord: https://discord.gg/KuHzF6zBPs Appreciate your time!
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
Launched alongside, November 2025
the whole month →
- IB
Life & fun · Nov 2025 · bitsnpieces.dev
- BBoing▲782
Life & fun · Nov 2025 · boing.greg.technology