Hopsule – Persistent memory and decision layer for AI development
Hi HN! I'm building Hopsule. If you use AI coding tools like Cursor, Copilot, or Claude, you’ve probably seen this happen: The AI writes good code - but it ignores your architecture. It doesn’t know: - why you chose a specific pattern - which conventions your team agreed on - which decisions are already locked in So it falls back to generic patterns, outdated examples, or random GitHub training data. Over time this slowly breaks the consistency of the codebase. Most teams try to fix this with: - giant Markdown files - wiki pages - long prompts - Slack threads But those aren't…
What it does
In the maker’s words, at launch
Hi HN! I'm building Hopsule. If you use AI coding tools like Cursor, Copilot, or Claude, you’ve probably seen this happen: The AI writes good code - but it ignores your architecture. It doesn’t know: - why you chose a specific pattern - which conventions your team agreed on - which decisions are already locked in So it falls back to generic patterns, outdated examples, or random GitHub training data. Over time this slowly breaks the consistency of the codebase. Most teams try to fix this with: - giant Markdown files - wiki pages - long prompts - Slack threads But those aren't machine-readable rules. So we built Hopsule. Hopsule turns architecture decisions into enforceable context that AI tools must follow. Example: Your team approves a decision: “All database access must go through the repository layer.” Hopsule records this as a rule and injects it into the AI context before code generation. No giant prompts. No manual context stuffing. No architecture drift. Website: https://hopsule.com Docs: https://docs.hopsule.com App: https://app.hopsule.com
Does the same job
all alternatives →- HPHopsule – Persistent Memory Layer for AI EngineeringMar 2026 · hopsule.com · ▲5
Hey Everyone, I'm neither the founder or affiliated with these guys. But when they showed me the product it really clicked a switch. I have been building products with AI since sonnet 4.0, and one of my issue, like many, consistency. Hopsule turns architecture decisions into enforceable context that AI tools must follow. Creates trackable, tasks which can be feed into your AI tools to do compound engineering. If you're building with Claude Code, Cursor, or Copilot. You can use their CLI or MCP.
- WAWe are building Git for dataJan 2026 · ▲9
Today you can easily adopt AI coding tools because you have git for branching and rolling back if AI writes bad code. We haven't seen this same capability for data and decided to build it ourselves. Nile is a new kind of data lake, purpose built for using with AI. It can act as your data engineer or data analyst creating new tables and rolling back bad changes in seconds. We support real versions for data, schema, and ETL. We'd love your feedback on any part of what we are building - https://getnile.ai/ What do you think?
- SRShikigami, run AI coding agents in parallel, each in a Git worktreeJul 2026 · shikigami.dev · ▲7
Hello all, I'm a software developer. Over the last few months more and more of my work has turned into using coding agents instead of typing the whole code myself. Usually a few claude sessions at once, sometimes codex, one per feature or per revealed bug. I ran them in a split terminal for a few weeks, and quickly spotted two main problems. The first is that I couldn't easily tell which agent was stuck waiting on me and which was still working, so I'd cycle through sessions and checking on them. The second one: agents sharing a single branch step on each other. Two of them could be editing…
- BTBrainGrid – The AI Product Planner: Structure Your Ideas for AIFeb 2026 · braingrid.ai · ▲5
Hi HN, I’m Nico, one of the co-founders of BrainGrid. We built this after spending the last year building software almost entirely with AI coding tools like Claude Code and Cursor. Writing code was fast. Finishing anything reliably was not. Coding was not the bottleneck. planning was. It was everything before the code: unclear scope, missing edge cases, poor sequencing, and vague intent. Once a project went beyond a single prompt, small changes started breaking unrelated parts of the app. In traditional teams, product managers and tech leads handle this. I was a Director of Product at Twilio…
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, March 2026
the whole month →
Switch from ChatGPT to Claude with import memory feature
AI · Mar 2026 · claude.com
