Tools Bench.

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

Last Brew Time: Jun 28, 2026, 7:24 AM PT

Insight

Builders aren't shipping standalone chatbots; they're wiring agents into the surfaces people already work in

01.AGENTS LIVING INSIDE EXISTING APPS

02.CODE-AND-KNOWLEDGE PLUMBING THAT FEEDS THOSE AGENTS

03.DOMAIN-SPECIALIST AGENTS PUT TO WORK

Featured

GitHub71.3K

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

A mature, institution-backed document-to-markdown engine that is among the most adopted tools of its kind and a near-default building block for feeding PDFs and Office docs into AI workflows. Backing by OpenDataLab, a steady release cadence, and consistent placement at the top of independent tool comparisons make this a strong project to track.

  • Roughly 71,000 stars with 6,000 forks, among the most widely adopted tools in this set by a wide margin
  • 175 releases with the latest (3.4.0) shipped June 2026, an unusually durable and active maintenance signal
  • Only 8 open issues against that scale points to tight, well-resourced upkeep rather than backlog drift
  • Backed by OpenDataLab (originated during InternLM pretraining), giving it real organizational durability
  • Independent 2026 comparisons repeatedly rank it a first choice for complex and CJK-heavy documents, strong third-party validation

feedbacks

What People Are Saying

  • "MinerU has the most GitHub stars for a reason - it's extremely capable at complex document layouts, especially CJK content"tech blog

  • "MinerU converts quickly and recognizes complex tables rendered via HTML"tech blog

  • "image handling may be less friendly, sometimes cropping images incompletely"tech blog

  • "highly recommended as a tool for pdf to markdown"DEVONthink forum

  • "parsing quality approaches commercial tools"tech blog

  • "Change OCR model?"GitHub discussion

HF Spaces16.6K likes

DeepSite v4

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.

youtube
DeepSite: This FREE Deepseek V3.1 AI Coder is INSANE!

Market Signal

Why It Has Market Pull

A vibe-coding platform that generates full applications from a prompt and has become one of the most-liked Spaces on Hugging Face. It is actively maintained with the developer responding to issues, runs on a stack of strong open models, and was promoted by HF's co-founders, making it a clear standout worth testing in real building workflows.

  • About 16,617 likes, among the very most-liked Spaces on Hugging Face
  • Reported 600k+ views and 15,000+ apps deployed within ~2 weeks of launch
  • Runs open models: DeepSeek-V3, Qwen3-Coder, Kimi-K2, GLM-4.x, MiniMax
  • Publicly promoted by HF co-founders Thomas Wolf and Julien Chaumond
  • Iterated through v1 to v4 in 2025; developer actively replies to reports

feedbacks

What People Are Saying

  • "I really like this project"Hugging Face community

  • "Is there a way to use older versions? These are hot garbage"Hugging Face community

  • "429 rate limit happens constantly during normal editing"Hugging Face community

  • "Small edits break the page... it deletes/rewrites sections"Hugging Face community

  • "Cannot Create a new Project or Run a prompt"Hugging Face community

  • "image in deepsite is broken"Hugging Face community

  • "can this run on mobile?"Hugging Face community

Hacker News373 pts

OpenKnowledge – open source AI-first alternative to Obsidian/Notion

Hi HN, Nick here. We’re launching OpenKnowledge ( https://openknowledge.ai/ ), a “what you see is what you get” markdown editor that has direct integrations with Claude, Codex, and other agents. Available as MacOS app or Web UI+CLI. Fully free/local and OSS. We built this because we wanted a Notion-like experience for writing and sharing markdown files across our team. Obsidian is the best alternative we tried, but found it doesn’t have a true WYSWIG UI and it didn’t integrate well with Claude/C... (373 points, 170 comments).

Market Signal

Why It Has Market Pull

OpenKnowledge is a local, open-source WYSIWYG markdown editor that bakes Claude/Codex/Cursor skills and an MCP server directly into the app, aimed at teams who want a Notion-grade editing feel over plain markdown files synced via git. It drew unusually strong attention and a packed comment thread for a launch, and comes from an established YC-backed team, which makes it a stronger signal among similar AI note-taking tools and worth a closer look for AI/content workflows.

  • Show HN reached 374 points with a very active discussion (June 25, 2026)
  • GitHub repo at 1.4k stars, 60 forks, only 7 open issues
  • Shipping fast: release v0.19.2 dated June 26, 2026; GPL-3.0, TypeScript
  • Built by Inkeep (YC W23), backed by Khosla Ventures and Y Combinator
  • Stronger product signal than most clones: built-in MCP server plus auto-installed agent skills, not just a markdown viewer

feedbacks

What People Are Saying

  • "I really wanted to like this, but unfortunately couldn't see how it improves my workflow"HN comment

  • "Don't take this the wrong way, but couldn't this have been a plugin?"HN comment

  • "I don't understand how Obsidian, a collection of markdown files, isn't already AI friendly."HN comment

  • "I just open the Obsidian folder in VS Code and BOOM, it is AI friendly."HN comment

  • "macos only? shame."HN comment

  • "I've decided to try daily driving this instead of Obsidian"HN comment

  • "Fully local ... an OSS Obsidian-like that syncs natively is an impressive accomplishment"HN comment

Hacker News27 pts

RLM-based local debugger for AI agent traces

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... (27 points, 10 comments).

youtube
Optimize an Agent End-to-End with Catalyst Tracing and HALO

Market Signal

Why It Has Market Pull

HALO ingests OTEL-compatible agent traces (Langfuse, Arize/OpenInference, or plain JSONL) and uses a Recursive Language Model to recursively decompose large trace sets, surface recurring harness-level failure modes, and feed fixes back into a coding agent. It shows one of the strongest signals among these tools, with a genuinely engaged technical discussion, fast GitHub traction, and published AppWorld benchmarks, making it well worth testing in real agent-development workflows.

  • Show HN on June 23, 2026: 27 points, 10 comments with substantive technical back-and-forth
  • GitHub repo around 1.1k stars, ~75 forks, 11 contributors (MIT, TypeScript and Python)
  • Featured on Trendshift as a trending repository; shared widely on social
  • Ships a desktop app, a pip-installable CLI, and a Python SDK; backed by inference.net / Context Labs
  • Reported AppWorld gains, e.g. Sonnet dev 73.7% to 89.5% and Gemini 3 Flash dev 36.8% to 52.6%

feedbacks

What People Are Saying

  • "Cool project... I'm curious how this compares to just using Claude Code directly and giving it a dump of the agent traces?"HN comment

  • "once you're at production scale the problem changes. You can't always fit 10,000+ traces in Claude Code and still have it be effective"HN comment

  • "I'm very curious to see a benchmark for this... extracting learning signal from deep-agent traces"HN comment

  • "There's some benchmarks in the repo for AppWorld. Looks promising"HN comment

  • "Very interesting. Haven't heard of RLMs before."HN comment

  • "when you say production-scale traces don't fit in context - does halo cluster failures first or just stream-summarize each trace?"HN comment

  • "What sort of systematic issues would teams typically uncover using HALO?"HN comment

Sources

GitHub

usestrix/strix

26.4K

Open-source AI hackers to find and fix your app’s vulnerabilities.

DeusData/codebase-memory-mcp

18.8K

High-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average repo in milliseconds. 158 languages, sub-ms queries, 99% fewer tokens. Single static binary, zero dependencies.

simplex-chat/simplex-chat

14.5K

SimpleX - the first messaging network operating without user identifiers of any kind - 100% private by design! iOS, Android and desktop apps 📱!

Robbyant/lingbot-map

8.0K

A feed-forward 3D foundation model for reconstructing scenes from streaming data

Product Hunt

Folio AI

322

Folio AI is the first real-time AI copilot for slides. It works directly in PowerPoint or Google Slides. And it's at least 6x faster than the competition, while being more performant.

QApilot's CoWork

304

CoWork turns existing test cases into executable mobile automation with AI planning, human-approved replanning, and real-device execution on iOS, Android, and Flutter.

Nada

244

Nada is the easiest way to turn your voice into music. Hum, sing, or whistle an idea, and instantly convert it into MIDI. Arrange your melodies on mobile, choose from a variety of instruments, and build songs wherever inspiration strikes. No AI generation involved, just your own creativity, captured faster.

RetroMac

198

RetroMac wraps your everyday macOS in the look and feel of classic computers: authentic CRT shaders over your whole screen, themed docks and desktops (Mac OS 9, Windows XP/98, BeOS, Amiga…), retro widgets and games, and a virtual webcam that pipes the CRT look into Zoom, Meet or Teams. Open source, free to use with an optional one-time upgrade for extra presets and webcam support.

Cloud World Model

172

Simulate AWS, GCP, Azure, OCI & DigitalOcean architectures to predict cost, performance, and resilience without provisioning real resources or paying a cloud bill. Built for learners practicing cloud skills and AI agents training on cloud optimization.

OpenBot

40

OpenBot is an open-source, local-first runtime that lets you drop diverse agents into a single workspace and orchestrate them using a clean tag-and-channel system. Just tag their capabilities and let OpenBot dynamically route tasks, sync collaborative workflows, and manage persistent threads right in your local file system. Fully modular, completely self-contained, and built for developers and makers who want absolute structural control over their agent fleets.

Hacker News

Adrafinil – keep a lid-closed Mac awake only while agents work

113 pts

A month ago there was a wave of posts and tweets about engineers walking around cafes and parks with their MacBooks propped half-open, as fully closing the lid forces sleep that stops their AI agents. Some people made snarky comments about using tmux or Amphetamine, and some defended their choice with “but I only need it sometimes, and forgetting to disable Amphetamine and finding my laptop discharged in my bag is worse.” This is a solution to this problem. Unlike caffeinate, it will prevent you... (113 points, 72 comments).

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

74 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... (74 points, 23 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).

Shumai – open-source Frame.io alternative for creative work

63 pts

Shumai is an open source platform for uploading creative files, managing projects, collecting precise feedback, sharing work, and collaborating with AI agents, all in one simple creative-first workspace. I’ve always liked the product design of Frame.io, and I wanted to build an alternative that feels just as polished, while being open source and easy to self host. You can deploy Shumai with docker compose in just a few minutes. It can also be installed from npm, though you'll need to provide you... (63 points, 6 comments).

Pure Effect – Reproduce production bugs on your laptop without a DB

58 pts

Hi HN, I think it's safe to say that the majority of developers don't give a second thought to writing code with I/O tangled in business logic. It's all too common to see code like: const user = findUser(email); if (!user) await saveUser(user); Now, you may ask: what's the big deal? When we write code like this, two things happen: 1. It gets harder to debug production bugs. Unless you have the exact same database and remote API services to connect to, you may fail to reproduce the bug. 2. You ha... (58 points, 14 comments).

HF Spaces

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...

Kolors Virtual Try-On

10.1K likes

Kolors Virtual Try-On is a Hugging Face Space tagged with gradio, region:us. It has 10118 likes on Hugging Face.

Wan2.2 Animate

5.1K likes

Wan2.2 Animate is a Hugging Face Space tagged with gradio, region:us. It has 5114 likes on Hugging Face.

MusicGen

5.1K likes

AudioCraft is a PyTorch library for deep learning research on audio generation. AudioCraft contains inference and training code for two state-of-the-art AI generative models producing high-quality audio: AudioGen and MusicGen. Installation AudioCraft requires Python 3.9, PyTorch 2.0.0. To install AudioCraft, you can run the following: We also recommend having ffmpeg installed, either through your system or Anaconda: At the moment, AudioCraft contains the training code and inference code for: MusicGen: A state-of-the-art controllable text-to-music model. AudioGen: A state-of-the-art text-to-sound model. EnCodec: A state-of-the-art high fidelity neural audio codec. Multi Band Diffusion: An EnC...

Arena Leaderboard

4.9K likes

Arena Leaderboard is a Hugging Face Space tagged with static, leaderboard, region:us. It has 4929 likes on Hugging Face.

The Ultra-Scale Playbook

3.9K likes

The ultimate guide to training LLM on large GPU Clusters Instruction to install and run locally Loading HTML fragments: There are two way to load HTML fragments: Compile them into html during build time Fetch them and insert them during run-time When to use what Use compile time fragments only on parts which you want to ensure are seen by every user right after page load (e.g logo) Use run-time fragments for everything else so that the final HTML is of reasonable size (<1MB idealy) How to add a new fragment Add it to the src/fragments folder (e.g. src/fragments/banner.html) For run-time fragments, add {{{fragment-name}}} to appropriate place in src/index.html (e.g. {{{fragment-banner}}}) For...

AI Tools — June 28, 2026 Edition | Agentic Brew