Ideas Lab

Real problems people complain about online, pulled every morning and scored out of 100. Build, validate, or skip. How scoring works

Ideas
4,145
BUILD
689
Sources
8
Last sync
43m ago
4,145 ideas

peerd – AI agent harness that runs entirely in your browser

Hey HN. http://peerd.ai is an AI agent harness that lives entirely in your browser as a web extension. You don’t have to install a separate “AI browser”. You don’t have to bolt on or run some external process or manage a clunky mcp integration. It’s just a fully contained web extension, written in no build vanilla JS with minimal non-browser dependencies, using your own provider keys, and Apache 2.This isn’t just a fun hack. While it has largely been a solo side project, I genuinely believe the browser and the web could be the most natural platform for AI agents to operate safely, autonomously, and most importantly without A2A middlemen (more on that in a sec). To demonstrate that point peerd doesn’t just drive browser automation. It spins up isolated sandboxes using tabs and worker instances to support various real workload types. Those include headless JS computational work, visual JS notebooks, personal client side apps, and real Linux VMs on top of wasm with full http networking.The industry discourse over the last several months has been dominated by “which substrate is the best for ai agent sandboxes” with many competing answers focused on different models and use cases. Cloudflare is one of the most prominent examples, positioning its v8 isolate based workers as the best in class solution thanks to faster than container startup times and strong isolation guarantees. The v8 isolate is of course the product of chromium, which runs on billions of browsers around the world for free. The browser as a whole is perhaps the most battle tested sandbox system in the entire software industry. It’s been built on 3 decades of learning from hostile content, hostile code, and hostile users. Native and cloud agents are necessarily rebuilding all or most of this posture from scratch. peerd doesn’t. It leverages everything the browser has to offer and pushes it to its functional limits, while inheriting its security baseline and isolation from the host system.Robus

Hacker News2mo agoToolAI

75FL score
0Sign in to vote

Write SaaS apps where users control where their data is stored

Hello HN,I would like to share with you linkedrecords.com - an open source backend as a service I'm working on since some time now. You can think of it as an firebase/convex alternative with an interesting twist.In 2018 I needed to write large software requirements/architecture documents in Google Docs. While I was annoyed by the limitations of Google Docs back then (no captions on figures, no automatic heading numbering, slow when docs are bigger,...) I was still fascinated by the real time collaboration features of it. So I've started a quest to understand how it works and I begun to implement an alternative to Google Docs.I was convinced that this kind of real time collaboration is the future so I've given it much thought how I could make this as generic as possible so I could use it in all future tools I would build.In the same time I was playing around with firebase (surprisingly you can not build a google docs alternative with firebase that easy as their real time collaboration does not provide merging text but rather just JSON). And back then I was also convinced that backend as a service is the right way to go. I was thinking that one of the most important reason we were still writing custom backend code is because of authorization.I also was faced with another problem when trying to make the backend as generic as possible: relations between entities are also domain specific. E.g. A Documents can have many comments.Luckily I was intrigued by another concept back in 2018 it was called web 3.0. Back in 2018 this had nothing to do with crypto. It was used as a term to refer to the semantic web and the resource description framework as one of its standards. There are also some RDF implementations which I could have reused but they are all XML and mostly Java based. I needed something light. Instead of implementing my own RDF product I took the idea of the RDF triplestore and came up with my own interpretation of it.Using concepts like: t

Hacker News3mo agoToolAI

65FL score
0Sign in to vote