Lobster is an initiative to make sure humans get the best out of an agent. The Lobster skin helps to get a thorough understanding and visibility of tasks they assign to agents.Join the waitlist @ https://getlobster.dev
So how it works, describe what should happen in plain language. Convoy extracts the screen, normalizes it to a platform-agnostic element table, asks Jev which control matches the intent, and taps it.
Btw iOS is tested, Android and Web is in test mode. Till then pls try using it and so share your feedbacks
Thank you!!!
I have a built a tool called Distill that converts any website to agent ready markdown/structure sub ~16ms. The tool works entirely local.Also did a benchmark on a corpus of 50 urls and Distill outperformed other competitors.Do give it a look: https://github.com/gokulnair2001/distill
Stop stitching together CLI tools. LLMForge handles the full pipeline — from model download to on-device deployment in one native window. No terminal. No cloud. Just results.
Hey HN,I built Crisper because every dictation tool I tried either sent audio to the cloud or gave me raw, messy transcripts I still had to fix.Crisper runs entirely on-device — no network calls, no account, no subscription. It does two things in sequence: transcribes using a speech model, then runs a local LM pass to strip filler words, fix grammar, and make the output sound intentional. The...
Product Healing Agent detects hidden UX friction and product gaps from real behavior, then tells your team what to fix first to improve conversion and retention.
Hi everyone, In today's age, creating application has been the easiest work to do, but when it comes to usage or adoption on a product people seem to be least interested. Isn't user adoption a vital thing? Why developers and companies are only interested in creating new features but not analysing the feedbacks or user traction?This purely comes from a genuine problem which I faced, and for...
Hi HN,I’m building a system that sits between application events and downstream consumers (analytics, decision systems, AI, etc.).The goal is to make raw product events more structured and canonical, so deriving patterns, insights, and higher level signals becomes easier.Before going deeper, I’m trying to sanity check the problem itself:* How do you validate whether a tool like this is actually...
Hi HN SwiftMock is a lightweight CLI tool that generates mock implementations for Swift protocols.The idea is to remove the repetitive boilerplate involved in writing mocks by hand, especially as protocols grow larger and start using async/await, throwing functions, and complex method signatures.The tool scans the codebase, identifies mockable protocols, and produces deterministic mock...
I’m building TTLI, a simple plug-in trust layer that any platform can add to detect fraud, score risk, and keep user trust intact without heavy inhouse systems.Curious if others here see value in an external trust layer or have thoughts on the challenges. Happy to chat with anyone interested.
I'm trying to understand how different teams handle platform integrity and user trust as they scale.For those running consumer apps, marketplaces, fintech products, or any platform with significant user activity: How do you currently build your “trust layer”?Specifically: -> How do you detect fake users, bots, device farms, or automated signups? -> What early signals do you rely on to...
The agenda is to, create a safer internet for every platform.
Your throughs?
I’m building a fast, simple, and private browser. No bloat, no tracking—just you and the web.What features would you love to see in a browser like this?If this sounds like your kind of browser, then join the waitlist: https://www.zerobrowser.co/
Hi HN,I’m building something I’ve always wanted but could never quite find—a browser that just gets out of your way. It’s called Zer0 Browser.The idea is simple: - No bloat - No tracking - Just the web, fast and cleanMost browsers today feel more like operating systems. They come bundled with shopping deals, crypto wallets, aggressive search suggestions, and background data collection. Zer0...
I’m building a new browser called Zer0 Browser, focused on being fast, minimal, and privacy first with absolutely zero bloat.
Most browsers today are packed with features users never asked for: ads, tracking, shopping suggestions, crypto plugins, and so on. Zer0 is the opposite—just a clean, focused experience for browsing the web.
Core Principles: Speed – Instant launch, lightweight UI, minimal memory usage
Simplicity – No tabs by default (like iA Writer for the web), clean reading mode, minimal UI
Privacy – No telemetry, no trackers, no background “smart” services
Exploring: Local LLM Integration – We're experimenting with optional on-device LLM support for things like page summarization, note-taking, and command-based browsing—all without sending your data to the cloud.
If this sounds interesting, I'm inviting early users to try the alpha version: https://www.zerobrowser.co/
Would love your feedback:
Would you use a browser like this?
What must-haves or deal-breakers come to mind?
What annoys you most about current browsers?
Thanks in advance!
1. Introduction: https://pledge.framer.website/home/introduction
2. Performance Test: https://pledge.framer.website/home/performance
3. Docs: https://pledge-gokulnairs-projects.vercel.app/documentation/...
4. Mission: https://pledge.framer.website/home/mission
Hello HN!I've developed Pledge, a lightweight and thread-safe reactive programming framework for Swift. It aims to simplify state management and event propagation in Swift applications by offering:Thread-Safe Observables: Ensure safe state management across different threads.Flexible Binding Mechanism: Easily bind observables to UI components or other data sources.Minimalistic API: Focus on...