11/14 Maker news - Yeogiseo Matna voting tool and AgentRun launch
Translated by AIView original
1. [New service] A web app that picks a meeting spot by link voting
I built this to cut down on the back-and-forth between KakaoTalk and map apps when picking a place to meet. You create a group, drop in links to restaurants or cafes, and the people you invited vote on each link to choose the spot. It supports Naver Maps, Daum Maps, CatchTable and Google Maps links.
Maker point Taking the flow of collecting several map links and voting on them as-is lets you quickly design the first feature of a small web app for meetups.
Source: GeekNews Show GN
2. [New service] A scanner that finds IP-conflicting devices by MAC
When several devices share the same IP, existing scanners sometimes only show one of them. This tool collects ARP replies and lists the different MAC addresses that respond on the same IP. Pick a device from the list and it pins the connection to that device for a while, so you can reach its admin page without unplugging cables.
Maker point Looking at how it collects ARP replies gives you a quick skeleton for your own tool that explores network devices.
Source: GeekNews Show GN
3. [New service] AgentRun, a self-improving agent platform
The GREP.AI team announced that they're open-sourcing AgentRun, which they grew inside their own agent harness, and growing it into a managed platform. They highlight agents that get better over time, workflows grounded in web sources, and multiple sandboxed execution environments.
Maker point If you're building a solo automation that hands multi-step work to an agent, you can borrow the structure of attaching sandboxed execution environments for testing from here.
Source: Hacker News Show HN
4. [Open source] TensorFold, an LLM inference engine for Metal, CUDA and Vulkan
An LLM inference engine written in Zig. It serves language models directly on Apple Silicon and NVIDIA GPUs, handling everything from reading checkpoints to tokenization and request processing. It's introduced as running without installing Python or MLX.
Maker point If you put a model on your Mac or GPU PC and have your app query it through a local API, you can try AI features on your own machine instead of an external API.
Source: GitHub Trending
What AI tools are you running on your own machine these days?
Comments 0
No comments yet
Be the first to comment.