Tools Bench.

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

Last Brew Time: Oct 13, 2026, 11:12 AM PT

Insight

Today's run shows builders routing around funded startups entirely

01.OPEN-SOURCE DEV TOOLING

  • **E2E – Open-source agentic testing framework for web

  • iOS and Android**

  • **Spotatui

  • a Spotify TUI/GUI without web API rate limits**

02.ONE-COMMAND SKILL INFRASTRUCTURE

03.THIN HUGGING FACE SPACE CLONES RIDING OTHERS' CHECKPOINTS

Featured

GitHub53.8K

cathrynlavery/diagram-design

Editorial diagram design for Claude Code, Codex, GitHub Copilot, Cursor, Factory Droid, and Pi. 44 diagram types. Self-contained HTML + SVG. No shadows. No Mermaid slop.

youtube
I Made 30+ Diagrams With One Claude Code Skill — No Figma

Market Signal

Why It Has Market Pull

This is a standout example of organic, explosive developer-tool adoption: a skill for AI coding agents that went from zero to over 50,000 stars in under six months, built by a credible repeat operator, with an active and diverse contributor base fixing real bugs rather than just starring the repo. It reads as a genuine category winner, not a flash-in-the-pan.

  • 53,951 GitHub stars and 3,424 forks, with a commit pushed the same day as this review
  • Hit #1 on GitHub Trending, gaining roughly 3,646 stars in a single day at peak
  • 39 open issues against a steady stream of closed, shipped fixes from multiple distinct outside contributors
  • Creator Cathryn Lavery previously bootstrapped a company to $55M+ in revenue before a private-equity sale, lending credibility to the launch

feedbacks

What People Are Saying

  • "Diagram-design is up to 550 stars on GitHub"X post

  • "fix(containment): honor important inline declaration priority"GitHub issue

  • "Standalone SVG export strips comment-like text inside quoted CSS selectors"GitHub issue

  • "fix(import): accept UTF-8 BOM in Mermaid sources"GitHub issue

  • "I tested Diagram Design, an open-source skill by Cathryn Lavery that turns Claude Code into a designer for diagrams"YouTube comment

  • "no shadows, no Mermaid slop"GitHub repo description

  • "GitHub's Top Trending Repo Today"dev blog post

GitHub34.3K

vercel-labs/skills

The open agent skills tool - npx skills

youtube
Como Usar las Skills de Claude mejor que el 99% de la gente

Market Signal

Why It Has Market Pull

This is a real, officially-backed product from Vercel Labs that has seen exceptionally fast builder adoption as a cross-agent package manager for AI agent skills, with strong early evidence of both scale and sustained engagement.

  • 34,402 GitHub stars and 2,947 forks (verified via GitHub API)
  • Official Vercel-backed ecosystem with its own changelog announcements and a companion skills registry
  • Independent coverage reports over 110,000 skill installs within four days of launch
  • Supports 75+ coding agents per the project's own docs
  • 1,109 open GitHub issues reflect an actively engaged user base rather than an abandoned project

feedbacks

What People Are Saying

  • "Private github repos not working?"GitHub issue

  • "skills.sh still renders stale SKILL.md after source update + fresh reinstall"GitHub issue

  • "Clone timeout when installing skills from large repositories"GitHub issue

  • "symlinking not working in cursor"GitHub issue

  • "frontend-design is recommended as the first skill to install — it improves every UI Claude builds from the first prompt"dev blog

  • "A single command works across Claude Code, Cursor, GitHub Copilot, OpenCode, and a dozen other AI coding tools"dev blog

  • "over 110,000 skill installs have already been recorded across 17 different AI coding agents"dev blog

Product Hunt299

Google EmbeddingGemma 2

EmbeddingGemma 2 is Google DeepMind's open embedding model that maps text, code, images, video and audio into one shared space, so you can search across them from a single query. At 740M parameters it runs on-device, needing about 567MB of RAM on a Pixel 11 Pro for the full multimodal model (191MB for text only). Apache 2.0 licensed, with weights on Hugging Face and Kaggle. MTEB Code score rises from 68.76 to 78.68.

youtube
EmbeddingGemma 2 is incredible

Market Signal

Why It Has Market Pull

EmbeddingGemma 2 is a genuine, credible release from Google DeepMind — a 740M-parameter, Apache 2.0, multimodal embedding model that runs in under 600MB of memory — and it already shows real builder adoption, not just press coverage. It is backed by one of the most trusted labs in the industry, has sparked independent community projects, and is being evaluated for production use in at least one widely-used open-source application.

  • 740M-parameter Apache 2.0 model released Oct 6, 2026 by Google DeepMind, built on the Gemma 4 architecture
  • 115,763 Hugging Face downloads in its first month, plus 71,914 for the community GGUF variant
  • Topped Product Hunt's featured apps list on Oct 12, 2026 per independent launch trackers
  • Immich, a widely-used open-source photo app, opened a formal proposal to adopt it as its Smart Search model
  • Runs across text/image/audio/video in under 600MB RAM on a Pixel 11 Pro, with support across MLX, vLLM, llama.cpp, Ollama, and LMStudio

feedbacks

What People Are Saying

  • "Someone on Reddit made a searching app that runs on EmbeddingGemma 2. Much better than MacOS's default Spotlight"X post

  • "[Feature] Add EmbeddingGemma 2 (sub-1B, multilingual, Apache 2.0) as a Smart Search model"GitHub issue

  • "Introducing EmbeddingGemma 2! Our lightweight, multimodal embedding model maps text, code, images, video, and audio into a single, unified embedding space"X post

  • "for embedding models in particular, it doesn't make sense to use a closed, proprietary, hosted-only model"HN post

  • "finally, a good moderate-size embeddings model — 270M for text-only, 440M total for text+vision"HN post

  • "why wasn't this compared to SigLIP 2?"HN post

  • "ran the demo locally and the exact classification example from Google's own demo failed, misclassifying a financial request with only 0.22 confidence"HN post

Hacker News9 pts

E2E – Open-source agentic testing framework for web, iOS and Android

Hey HN, I’m Oskar, one of the co-founders of TesterArmy. We built an agentic testing framework for web and mobile because traditional testing frameworks can’t keep up with the pace of how we build software today. e2e allows you to mix deterministic and agentic APIs in the same test. You can write the test cases using familiar APIs like `getByText()`or `getByTestId()` which you can also mix it with agentic `agent.act()`. The unique part about it is that e2e will cache and replay the agentic actio... (9 points, 1 comments).

youtube
e2e - Agentic End-to-End Testing (Full Tutorial)

Market Signal

Why It Has Market Pull

E2E, built by YC-backed TesterArmy, is a rare case of a genuinely viral open-source launch backed by real funding and real paying usage: it gained over 1,000 GitHub stars within 24 hours, its launch post pulled roughly 449K views, and well-known engineering leaders publicly endorsed it. The underlying company has raised institutional capital from Y Combinator and notable angel investors and already counts recognizable companies among its customers.

  • Over 1,000 GitHub stars gained in under 24 hours; repo has grown past 9,900 stars overall and trended on GitHub
  • TesterArmy raised $1.2M pre-seed from Y Combinator, AIP Seed, Eight Capital, and Script Capital, with angels including Guillermo Rauch (Vercel) and Walden Yan (Cognition)
  • Launch post reached about 449K views on X
  • More than 50 companies reportedly already use the underlying product, including Resend, bolt.new, Novu, CodeCrafters, and Rork
  • The row's own Hacker News Show post landed only 6 points/0 comments despite the much larger traction elsewhere

feedbacks

What People Are Saying

  • "This looks great"X reply

  • "The agentic mobile testing game just changed"X reply

  • "We've all been building our version of this, but this is better. Well done!"X reply

  • "it mixes deterministic and non deterministic testing, along with caching actions so you don't have to call the model multiple times"X reply

  • "His launch post pulled about 449K views"X post

  • "Trying e2e: Write End-to-End Tests with Plain English and Real Assertions"Dev.to article

Sources

GitHub

morluto/rea

199.6K

Reverse engineer anything with agents, from app behavior down to native binaries.

debpalash/VoiceStudio

62.5K

VoiceStudio is the open-source, fully-local ElevenLabs alternative — voice cloning, voice design, video dubbing, dictation, transcription & audiobook creation in 646 languages.

bilawalsidhu/gods-eye-view

55.0K

A spy satellite simulator in your browser, except the data is real. Live open source spatial intelligence on a photorealistic 3D globe.

mksglu/context-mode

28.0K

Context window optimization for AI coding agents. Sandboxes tool output (98% reduction), persists session memory, and enforces routing across 17 platforms via MCP + hooks.

storytold/artcraft

24.9K

ArtCraft is an intentional crafting engine for artists, designers, and filmmakers

Product Hunt

ngram 2.0

261

Turn ideas, documents, and recordings into narrated videos. ngram analyzes uploaded videos and screen recordings to identify key moments and build clear product walkthroughs. Create animated explainers or add an AI avatar presenter, then refine your video in chat.

Base Code

197

Base Code lets teams work directly on existing codebases from the browser. Connect GitHub, describe changes in plain language, preview them live, and send completed work through standard pull requests.

guilds.run

193

Open-source AI agent teams you can actually inspect. Debug every prompt, tool call, memory change, and agent interaction. Control permissions, edit context, and choose the model for each agent. Self-host for free. Stop guessing what your agents are doing.

Manifold MCP

157

Manifold gives your agent live data from every marketing channel, in one hosted MCP server for Claude, ChatGPT, Cursor and Claude Code. Find keywords one push from page one. See if ChatGPT and Gemini recommend you. Watch the TikToks winning your niche. Read your competitors’ live ads. Catch Reddit threads asking for what you make. Get verified emails for buyers looking right now. Ask in plain words and your agent does the work. No pile of tabs. Free to start.

Ancher 2.0

151

Ancher is an AI knowledge workspace that turns everything you read, watch, and save into work you can actually use. Capture articles, videos, PDFs, posts, and ideas without organizing folders. Ancher connects related knowledge, lets you search and ask questions across your sources, and helps turn what you've learned into research briefs, reports, presentations, and content in your own voice. Your saved knowledge becomes more useful every time you add to it.

Hexis Cloud

124

Hexis is one company brain for your team and all its AI agents. Through Hexis, your team can align on shared context that's important to align on, non-technical teams can work and collaborate on skills and your team can have hundreds of tools accessible via one MCP gateway. All the information is owned by someone, meaning it's context you can trust, and Hexis enables access control to a single file. It only takes a few minutes to set up, and is free for up to 3 people. Would love your feedback!

YC Launch

Native: Artificial intuition for marketing

2

Native predicts what you should post, and then runs it across your socials and ads. Native · Fall 2026 · B2B Tags: Artificial Intelligence, Machine Learning, Social Media. Website: https://native.no/

Hacker News

Breadcrumb, record everything on your mac + context manager for AI

49 pts

Hi HN, I'm Justin. Breadcrumb records everything you do on your Mac (screen + meetings + AI transcripts + what you and your AI decided) and turns it into memory your AI can search. It's local and encrypted. You can also teach it rules by talking to it and it makes sure the right rules turn up in the right context. Works with Claude Code / Codex / Cursor / opencode. All of this is exposed to your AI as 30+ MCP tools (here's the definitions): https://innerloop.works/breadcrumb/mcp I started it in... (49 points, 8 comments).

Spotatui, a Spotify TUI/GUI without web API rate limits

4 pts

About 11 months ago i revived the amazing but unfortunately unmaintained spotify-tui. Since then I have been working on it almost non stop, finding a lot of problems on the way. One of the main issues I ran into were the Spotify web API rate limits. Third party clients like spotatui are dependent on a developer app made in the Spotify developer dashboard, which get rate limited a lot and Spotify has slowly been deprecating endpoints from it. The workaround for this used to be using a shared clie... (4 points, 0 comments).

Proton Drive for Linux

122 pts

Hello everyone, I wanted to share this small project that I have. It's born out of a necessity that I had, because Proton doesn't ship a Linux version of the drive yet (it's apparently coming out later this year but who knows?) and all the current solutions are hacky and wonky at best. But for the time being I really needed to mount my Proton drive on my Linux machine the way I mounted it on my Mac, so I (and my faithful AI slave) did this small utility that allows you to mount your Proton Drive... (122 points, 45 comments).

Opus 5.5 made a fully featured Figma clone in 3 hours / 8 prompts

39 pts

I think a chasm has been crossed. Opus 5.5 few shotted a fully working Figma for me. In the past year I have been trying to benchmark vibe coded apps and see how good they get and this week it did it with very very little guidance from me. I basically just told it reach feature parity. Prior to this vibe coded apps would resemble the real software, this is not a resemblance it's the real thing. As a casual Figma user in the past, I can't really see a difference. And if there is one you have an L... (39 points, 43 comments).

MemorySync – Memory for AI Agents

22 pts

I created MemorySync because I kept running into the problem of AI agents not being able to remember useful things about users between conversations. I wanted an easy way to give the agents memory in my application without having to build the whole thing myself. MemorySync is an API for giving AI agents memory. The application sends MemorySync information provided by the user, MemorySync saves useful memories for that user, and they can be retrieved later when needed. Give it a try without creat... (22 points, 1 comments).

HF Spaces

Jev Decision Index

628 likes

Benchmarks and news on various repros of TypeSafe's Jev Who is rebuilding TypeSafe's Jev (System One / RLCD) in the open? This static Space opens on the Decision Index leaderboard; the News tab tracks the artifacts in one combined grid, color-coded by kind: Decoding: parallel constrained decoding on stock models (inference technique, no new weights) Diffusion: text diffusion models run in a "Jev mode" Trained: Jev-like scoring heads and fine-tunes, weights often on the Hub, promised models listed last Prior art: "this already exists" claims Explainers: architecture speculation, explainers, benchmarks and roundups Cards sort by a trending score: ♥ likes on X + 5 × GitHub stars + 8 × Hub likes...

LightOnOCR-3 Demo

45 likes

OCR to Markdown with layout grounding Demo of LightOnOCR-3: drop a page, get Markdown with layout grounding. LightOnOCR-3 Demo is a Hugging Face Space tagged with gradio, region:us. It has 45 likes on Hugging Face.

MiniMax H3 Turbo LoRA

1.5K likes

Video generation with a synchronized soundtrack MiniMax-H3 — unquantized, split across two Spaces Joint video and soundtrack out of a single denoising pass, at bfloat16 with no quantization anywhere. This Space is the denoising half: the 61.73 GiB transformer and the two autoencoders. The 62.14 GiB Qwen3-VL conditioner runs in qwen3vl-conditioner, which this Space calls over the gradio API for every request. The weights are the public MiniMaxAI/MiniMax-H3 diffusers checkpoint. MiniMax-H3 is 195.9 GiB in bfloat16 and a ZeroGPU Space is evicted at 150 GB of storage. An unquantized single Space is therefore impossible, which is why quantized demos of it run NVFP4 or float8 weights. Cut the Mini...

Wan2.2 14B Preview

495 likes

generate a video from an image with a text prompt Wan2.2 14B Preview is a Hugging Face Space tagged with gradio, mcp-server, region:us. It has 495 likes on Hugging Face.

QIE-2511 Rapid-AIO LoRAs Fast (Experimental)

445 likes

Demo of the Collection of Qwen Image Edit LoRAs QIE-2511 Rapid-AIO LoRAs Fast (Experimental) is a Hugging Face Space tagged with gradio, mcp-server, region:us. It has 445 likes on Hugging Face.

Qwen Image 2.1 Uncensored GGUF Studio

415 likes

Qwen-Image-2.1 Uncensored GGUF demo on ZeroGPU Qwen-Image-2.1 Uncensored GGUF Demo A Gradio application running KasugaiSakura/Qwen-Image-2.1-Uncensored-Abenzerps-GGUF on Hugging Face ZeroGPU (zero-a10g). Highlights Text-to-Image ("Create an image"): High-resolution image synthesis from complex prompts. Image Editing ("Edit an image"): Upload a reference image and describe edits to apply. Transparent PNGs ("Transparent PNG"): Generate sticker/cutout assets with native alpha channel. ZeroGPU Fast Sampling: Uses qwen-image-2.1-UC-Q4KM.gguf packed into BF16 at startup for instant GPU dispatch without runtime dequantization overhead. Metadata Embedded: Downloaded PNG images contain generation par...