Tools Bench.

Product launches and open-source repos with enough signal to earn a second look.

Last Brew Time: Jun 25, 2026, 7:38 AM PT

Insight

The throughline this run is agents that no longer just answer but act, handed a wallet and a mouse

01.SINGLE-USE CARDS AND PAY RAILS FOR AUTONOMOUS SPEND

Buy by Agentcard
You can now buy Doordash from Claude.
Stripe.Directory
Stripe Directory is a single discovery layer for developers and AI agents to find businesses on Stripe.

02.NATURAL-LANGUAGE CONTROL OF LIVE WEB INTERFACES

03.CRM AGENTS THAT FIRE ON A SCHEDULE WITH NO HUMAN CLICK

Featured

GitHub69.2K

opendatalab/MinerU

Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.

youtube

PDF转markdown，AI知识库绝配，科研文献整理必备，MinerU将PDF转化为机器可读格式的工具

Market Signal

Why It Has Market Pull

An institutionally-backed document-to-markdown/JSON engine (PDFs, Office docs) for LLM and agentic workflows, from OpenDataLab (Shanghai AI Lab). It is the most established and credible tool in this group, with deep engineering history and heavy production use.

~69.3k GitHub stars, supported by years of history
Backed by OpenDataLab; reported production use by Google, Huawei, Alibaba and 100+ enterprises
Mature release cadence: 175 releases, latest v3.4.0 (~June 18, 2026), with thousands of issues and discussions
Benchmarked in independent 2026 PDF-to-markdown roundups against Marker, Docling and others
License recently migrated from AGPLv3 to a custom Apache-2.0-based open-source license

feedbacks

What People Are Saying

"parsing quality approaches commercial tools; best for CJK documents and complex academic papers"Comparison roundup
"MinerU takes 3+ minutes before it actually starts processing a 120-page OCR PDF"GitHub issue
"video memory is still occupied and not released after the task completes"GitHub issue
"CUDA out of memory; tried to allocate 110 MiB"GitHub issue
"vLLM speed far below the official docs"GitHub issue
"Arabic, Hindi and Urdu were not reliable for text extraction in testing"Tutorial review

GitHub19.7K

alibaba/page-agent

JavaScript in-page GUI agent. Control web interfaces with natural language.

github

Page Agent — in-page GUI agent: control web interfaces with natural language

Market Signal

Why It Has Market Pull

An official Alibaba JavaScript in-page GUI agent that controls web interfaces via natural language, installable via a bookmarklet. It shows some of the strongest organic third-party traction in this group, with a lively Show HN debate and independent hands-on reviews.

~19.7k GitHub stars since early March 2026
Official alibaba organization repo, MIT-licensed, with a polished bilingual docs site
Very active: 33 releases, latest v1.10.0 (June 15, 2026), and a busy issue tracker
Show HN scored 147 points / 76 comments; independently reviewed with a hands-on test
One field test reported a 30-field form dropping from ~3 minutes to ~35 seconds with the agent

feedbacks

What People Are Saying

"Very cool! I'm particularly impressed by the bookmark trick to install it on a page."HN comment
"How is this secure? It could impersonate the user pretty easily."HN comment
"I set the LLM to a local endpoint and the extension crashed at startup."HN comment
"Instant 'Task aborted' error when executing tasks via MCP due to a re-render"GitHub issue
"Cannot use a custom large model"GitHub issue
"an operation that took 3 minutes now takes 35 seconds"Hands-on review blog

GitHub18.4K

google-labs-code/design.md

A format specification for describing a visual identity to coding agents. DESIGN.md gives agents a persistent, structured understanding of a design system.

Market Signal

Why It Has Market Pull

An official Google Labs spec (tied to Google Stitch) for giving coding agents a persistent, structured understanding of a design system. Strong institutional backing and rapid adoption, though much of its early momentum is driven by Google's own promotion and the format is still settling.

~18.5k GitHub stars in roughly two months
Published by the official google-labs-code organization under Apache-2.0; distributed on npm as @google/design.md
Active release cadence: 4 releases, latest v0.3.0 (June 15, 2026), with the token schema still evolving
Featured on Google's official blog and amplified across developer social feeds
Still alpha-stage: the spec format is actively changing, so expect breaking adjustments

feedbacks

What People Are Saying

"DESIGN.md is very useful for quickly getting something up and running"HN comment
"Very useful. Stealing it."HN comment
"how is DESIGN.md meant to play nicely with Figma tokens or tailwind.config.js?"HN comment
"codifying design rigidly in docs may slow you down more than it's worth"HN comment
"everything defined here is already in my themes config or component files"HN comment
"npm install @google/design.md failing"GitHub issue
"accessibility considerations for AI-generated UI"GitHub issue

Product Hunt552

Tencent EdgeOne Makers

Tencent EdgeOne Makers is an edge platform for modern web apps and AI agents. Build with your preferred frameworks and deploy through familiar CLI, Git, and CI/CD workflows. Get built-in agent runtime, sandboxed tools, memory, observability, model gateway support, serverless functions, and storage—without stitching together complex infrastructure. Add AI agents to existing products or launch new AI applications in minutes. Deploy AI agents like web apps.

youtube

从0到1开发并部署上线第一款生产级Agent #edgeone #ai #ai编程 #vibecoding

Market Signal

Why It Has Market Pull

A real, well-resourced product from Tencent Cloud that topped its Product Hunt launch day with 552 votes and reportedly serves 300,000+ users. It positions edge hosting plus a built-in agent runtime (memory, sandboxed tools, observability, model gateway) as a single deploy target, with framework-agnostic support for the Claude SDK, OpenAI SDK, LangGraph and CrewAI. Strong launch momentum and credible backing make it worth a closer look.

Reached #1 Product of the Day on Product Hunt (June 24, 2026) with 552 votes
Reported 300,000+ users since the December 2025 launch
Backed by Tencent Cloud's 3,200+ global edge nodes for built-in CDN acceleration
Companion deploy-tool repo carries ~423 GitHub stars / 57 forks, MIT-licensed and actively maintained
Works natively with Next.js, Claude SDK, OpenAI SDK, LangGraph and CrewAI with no proprietary SDK lock-in

feedbacks

What People Are Saying

"Migrating from CF Pages was quite simple"Dev blog review
"Finally, an unlimited-traffic free CDN with domestic acceleration nodes"Dev blog review
"many features are still missing, for instance it doesn't support Hugo's automatic deployment"Dev blog review
"many advanced features are removed, like the header-based cache-key customization I wanted most"Dev blog review
"Might be a good choice for pages aimed at visitors inside China"Self-hosting forum
"free for 14 days and a massive 10GB"Self-hosting forum

Product Hunt128

FUTO Swipe

FUTO Swipe is a family of small, open models for accurate swipe typing. It includes a layout-agnostic encoder, a layout-specific decoder, and a lightweight context language model. The full system runs efficiently on-device with a very small footprint, and FUTO has also released the 1 million swipe dataset used to train it.

youtube

FUTO Keyboard Updates: The Swipe Dataset is Public!

Market Signal

Why It Has Market Pull

FUTO Swipe is a family of tiny open swipe-typing models from FUTO, the established team behind FUTO Keyboard and Grayjay, shipped fully on-device with a publicly released 1-million-swipe training dataset. It drew strong cross-community engagement on launch and rides the credibility of FUTO Keyboard's existing repo and millions of installs, positioning it as a privacy-first rival to Gboard and SwiftKey.

Launched June 24, 2026 on Product Hunt with 128 upvotes, reaching the daily top 10
Parent FUTO Keyboard repo carries ~3.4k GitHub stars with active releases
Open release includes all 3 models (635K encoder, 300K decoder, 1.5M context LM) plus the full 1M-swipe dataset
Reported ~4% top-4 fail rate with on-device inference in milliseconds
Backed by FUTO's grant-funded open-source org rather than venture money

feedbacks

What People Are Saying

"Releasing both the models and the 1M-swipe dataset is what stands out here."Product Hunt comment
"Privacy-first UX only works when the core interaction feels good, and this is a rare case where it does."Product Hunt comment
"The swipe demo on the website is neat; I could even use it with my mouse."HN comment
"it feels as good as the Google keyboard"HN comment
"the best free and private keyboard I found"HN comment
"Can only do one language at a time versus Gboard's simultaneous support."HN comment
"random capitalization and odd next-word suggestions"HN comment

Sources

GitHub

garrytan/gstack

115.5K

Use Garry Tan's exact Claude Code setup: 23 opinionated tools that serve as CEO, Designer, Eng Manager, Release Manager, Doc Engineer, and QA

NanmiCoder/MediaCrawler

52.6K

小红书笔记 | 评论爬虫、抖音视频 | 评论爬虫、快手视频 | 评论爬虫、B 站视频｜评论爬虫、微博帖子｜评论爬虫、百度贴吧帖子｜百度贴吧评论回复爬虫 | 知乎问答文章｜评论爬虫

apple/container

42.9K

A tool for creating and running Linux containers using lightweight virtual machines on a Mac. It is written in Swift, and optimized for Apple silicon.

JCodesMore/ai-website-cloner-template

20.0K

Clone any website with one command using AI coding agents

xbtlin/ai-berkshire

1.6K

AI 时代的伯克希尔：基于 Claude Code 的价值投资研究框架。巴菲特·芒格·段永平·李录四大师方法论 + 多Agent并行研究。| AI-era Berkshire: a value investing research framework built on Claude Code. 4 masters' methodologies + multi-agent adversarial analysis.

aws/agent-toolkit-for-aws

1.0K

Official, AWS-supported MCP servers, skills, and plugins to help AI agents build on AWS

Product Hunt

Propane

509

Propane gives your product team and agents one connected, always-current view of your customers. Automatically collected from all your tools. Collaborate on a shared canvas. Commit straight to any coding or design agent. Secure, maintained, always on. You just build products people love.

Stripe.Directory

280

Stripe Directory is a single discovery layer for developers and AI agents to find businesses on Stripe—including Stripe Apps, Projects.dev providers, and mpp.dev services—and integrate them without the need to manually search for them.

Customer Relationship Agents by Clarify

200

Agents that handle the CRM work you've been doing manually — pipeline digests, lead enrichment, data hygiene, call coaching, and much more. They run on a schedule or a signal, work across your stack, and fire without you touching them. Set them to run fully autonomously or require your approval before acting. Describe what you want to automate or start from a template. Clarify builds it from there.

Mindstone Rebel

185

Rebel is a desktop AI workspace for agentic work. It connects your memory, meetings, files, actions, automations, and tools so AI agents can help with real work — while keeping sensitive actions behind approval checks. Built Fair Source, with portable workflows and model choice.

Buy by Agentcard

172

you can now buy Doordash from Claude. connect Agentcard to your Openclaw, Claude or whatever you use, ask for what you want and we handle the rest. it feels magic. install from agentcard.sh and make your first order for free!

React UI Kit V7

121

Build chat, voice, and video experiences in React faster with CometChat React UI Kit V7. Get production-ready messaging components, calling, AI features, theming, localization, and customization out of the box. New in V7: a more flexible component architecture, plugin-based extensibility, improved customization, and support for React, Next.js, React Router, and Astro. Ship a fully featured chat experience effortlessy.

Hacker News

We post-trained a model that pen tests instead of refusing

93 pts

Anthropic and OpenAI's publicly available models are explicitly guard-railed so that they refuse offensive tasks. And their cyber-focussed models are gated for enterprises. This leaves SMEs and mid market open to major vulnerabilities. AI can be used as both an adversarial and defensive tool in the world of cyber. A worst case outcome is if only the adversaries have access. Meanwhile, most existing AI cyber tools are just wrappers. The problem is that they still have all the guardrails on from t... (93 points, 40 comments).

peerd – AI agent harness that runs entirely in your browser

63 pts

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... (63 points, 22 comments).

Selector Forge – browser extension for AI-generated resilient selectors

38 pts

Hi HN, I'm Ahmad from the Intuned ( https://intunedhq.com ) team. Today, we're releasing and open-sourcing Selector Forge ( https://selectorforge.ai/ ), a browser extension that generates reliable CSS/XPath selectors using AI. You can use it to create a selector for a single element or for an array of elements. The selectors it creates are meant to be "semantic" and more resilient to page changes than what Chrome DevTool’s “Copy Selector” (and other similar extensions) give you. Those tend to ha... (38 points, 2 comments).

RLM-based local debugger for AI agent traces

25 pts

We built HALO (Hierarchal Agent Loop Optimizer), an open-source tool for debugging and optimizing AI agents using their execution traces. It’s a loop. Run your agent, feed the traces to HALO, get the report, apply the fixes, then re-run your agent. HALO takes in OTEL compliant traces from AI agents using tracing frameworks such as Langfuse, Arize/OpenInference, or even just plain JSONL. It uses an RLM (Recursive Language Model) to more efficiently break trace analysis into smaller subproblems in... (25 points, 10 comments).

Ito – Code reviews that run code

10 pts

I'm Evan and I made Ito.ai It's code review that actually runs your code. The result is that it finds more bugs with a smaller false positive rate. You get screenshots, videos, and run logs. I built Ito because manual verification sucks and I thought I could automate it. I've been a professional developer for 15 years and now more than ever, I'm just spending time making sure things work. I looked into all of the other agentic testing tools that click around a url you give them and frankly, they... (10 points, 8 comments).

Agnes AI – Free multimodal API (text, image, video), OpenAI-compatible

8 pts

Hi HN, I'm Daniel, part of the team at Agnes AI, a Singapore-based AI lab. We've been building quietly for a while and I wanted to share what we've made with this community and get honest feedback from people who actually build things. What is Agnes AI? Agnes AI gives you access to three models through a single free API. We have text, image and video models. The API is fully OpenAI-compatible, so you can just swap the base URL to https://apihub.agnes-ai.com/v1 and use your Agnes API key. No othe... (8 points, 3 comments).

HF Spaces

DeepSite v4

16.6K likes

Generate any application by Vibe Coding it DeepSite is a Vibe Coding Platform designed to make coding smarter and more efficient. Tailored for developers, data scientists, and AI engineers, it integrates generative AI into your coding projects to enhance creativity and productivity. DeepSite v4 is a Hugging Face Space tagged with docker, region:us. It has 16617 likes on Hugging Face.

AI Comic Factory

11.1K likes

Create your own AI comic with a single prompt Last release: AI Comic Factory 1.2 The AI Comic Factory has an official website: aicomicfactory.app For more information about my other projects please check linktr.ee/FLNGR. If you like the AI Comic Factory, let me know! I am always creating new spaces and exploring new ideas for demos, meaning I don't have much time to take care of all of them (I wish I could clone myself or ask robots to do it). If you appreciate the AI Comic Factory and would like to leave a tip, that would be very kind 🫶 First, I would like to highlight that everything is open-source (see here, here, here, here). However the project isn't a monolithic Space that can be dupli...