PH热榜 | 2026-02-07

These are so cute :) Wishing good luck with the launch :)

This is dope for sparking ideas. The hard part is what happens after the notes. How do you move from brainstorming → real execution? We’ve been building ImpactOS to help teams turn ideas into concrete action steps. Would love to hear how folks here handle that handoff.

Nice idea! Congrats and good luck with the launch!

Combining a card deck for brainstorming with a notes app is a unique approach to creative blocks. The UI looks very clean. Is there an offline mode or is it purely cloud-sync based?

When can we expect it for Android users?

When Android folks are deprived from amazing apps like InspireNote, here's how we feel, hehehe...

You’ve done an excellent job, sir!

I’ve Checked Obooko and got some feedback for you.

It would be better to add a Google login or use third-party services like Clerk.

Since I have a design background, the Brand Sprit I see in Obooko is quite similar to a website like Patreon, and it fits perfectly. However, you can pursue what they’ve already done and enhance the visual and user experience.

Yeah! Let's stop dumb-scrolling. It's a super cool alternative to make our time not only more productive but also healthier. Happy to see you helping on this George. All the best here!

What about the commercial titles that should normally be purchased in the store? Those are excluded?

I like the idea. I am trying to read more again (but paper books), because when I have a device nearby, I am so tempted to use it.

Congrats on rebuilding Obooko and the launch!

I'm curious tho about the experience users have regarding long-form reading mixed with ads tho. While I agree that it's a great way to keep the platform free, I find it hard to find it similar to YouTube's model - when it comes to video, it's easier to handle an ad every few minutes because the content is short.

But if I'm deep into a book for hours, do ads break the flow? How did you balance monetization with not killing the reading experience?

Love the 'YouTube model for books' idea - readers get free content, authors still earn. That's a win-win that could actually scale. The classics collection sounds great too, might finally finish Wuthering Heights.

Hey Product Hunt! 👋

I'm Rohit, the maker of SkillKit.

The problem I kept running into: My skills for Claude Code were useless in Cursor. My team was on Copilot. Each agent has its own format. Worse: every time I close a session, my AI forgets everything it learned.

So I built SkillKit - a universal platform for AI coding agents that goes way beyond skill management.

npx skillkit@latest

What makes it different from "skill installers":

🔄 Cross-Agent Translation (32 agents)

skillkit translate my-skill --from claude --to cursor,codex,copilot

One command. Write once, use in Claude Code, Cursor, Codex, Windsurf, Copilot, Gemini CLI, and 26 more.

🧠 Session Memory That Persists

Your AI agents learn patterns, but that knowledge dies every session. SkillKit captures learnings with semantic embeddings and makes them persistent:

skillkit memory compress    # Extract patterns from sessions

skillkit memory search "auth patterns"  # Recall past learnings

skillkit memory export      # Turn learnings into shareable skills

🤖 Multi-Agent Team Orchestration

Spawn teams of AI agents with leader/teammate hierarchies, task assignment, plan approval workflows, and code review stages:

skillkit team init

skillkit message send       # Inter-agent messaging

skillkit workflow run       # Orchestrate multi-step tasks

🌐 Mesh Network for Distributed Teams

Your agents can communicate across machines with E2E encrypted P2P:

- Ed25519 cryptography & XChaCha20-Poly1305 encryption

- UDP multicast LAN discovery

- Trust management with fingerprint verification

🎯 AI-Powered Recommendations

skillkit recommend --explain
# > 92% match: vercel-react-best-practices (Next.js detected)
# > 87% match: tailwind-v4-patterns (Tailwind 4 in package.json)

📚 Built-in Methodology Packs

TDD, Design-First, Root Cause Analysis, Structured Review - battle-tested development methodologies baked in.

More Features:

- 🧪 Skill Testing Framework with assertions

- 🔧 Auto-generate CI/CD configs (GitHub Actions, GitLab CI)

- 🌳 Hierarchical skill taxonomy with tree navigation

- 📡 Self-host your skills (RFC 8615 well-known URIs)

- 🔌 Plugin system for extensions

- 📊 Quality scoring and security audits

This is for you if:

- You switch between AI coding agents

- Your team uses different tools

- You want your AI to actually remember and learn

- You're building multi-agent workflows

- You want enterprise-grade skill management

Fully open source.

Website: agenstskills.com | Docs: agenstskills.com/docs

Really cool to see this direction. The portability issue between Claude Code, Cursor, Copilot, etc. is something I’ve run into as well — each agent having its own “skill dialect” makes it hard to build consistent workflows.

I’m working on a different layer of the stack (Iceberg Framework), where the focus is on deterministic execution and validation for LLM‑powered systems. What you’re doing with Skillkit around persistent skills and cross‑agent translation actually complements that nicely — one solves the portability problem, the other solves the reliability/consistency problem.

Love seeing more tooling appear in this space. The ecosystem really needs it.

Package manager for AI skills is brilliant! I'm working with Claude Code and the problem of 'teaching' it project-specific patterns is real. Does Skillkit let me create custom skills like 'how we structure Phoenix LiveView components' and have the agent remember across sessions? The cross-platform support (Claude, Cursor, etc.) is the killer feature - skills shouldn't be tool-locked. How does versioning work? Congrats on open sourcing this!

The cross-agent translation for 32 agents is impressive. I use Claude Code daily and switching context between tools is painful. Curious about the Primer feature — how does auto-generating instructions work in practice? Does it analyze your codebase patterns and suggest skill definitions, or is it more template-based? Also, for team scenarios where different members use different agents (Claude Code vs Cursor vs Copilot), how does skill synchronization work across the mesh network?

Same boat as kxbnb - been manually converting CLAUDE.md to .cursorrules whenever switching tools. The translate command alone is worth it. Also curious about the mesh network for multi-agent setups, sounds like it could be interesting for coordinating different specialized agents.

The concept of a package manager for AI skills is brilliant. As a dev, I’m curious—how does it handle dependency versioning for different LLM models? Great work on making it Open Source!

I maintain about 15 Claude Code skills for my daily workflow and the portability problem is real. Right now if I want the same behavior in Cursor I'm manually rewriting CLAUDE.md into .cursorrules. The translate command would save me a lot of time.

Curious about the memory feature -- how does it handle conflicting patterns across projects? Like if I have one repo that uses snake_case and another that uses camelCase, does it scope learnings per-project or is it global?

Love this, its really smart idea. Live captions at events are always either overcomplicated or crazy expensive, so a simple browser link solution makes a lot of sense. Great launch, Martin & Jarek!

Love the friction-less 'browser-first' approach—I actually shared that same philosophy when I speed-ran the build for Featmap.app on a 4-hour train ride to create a completely free feedback board that founders can set up in 30 seconds.

Awesome project! Congrats on the launch! 👏🏻

This is very dope! Congrats on the launch!

A simple solution for a very frequent problem, especially if you present to large audiences. I'm excited to see how you guys build this out :)

Hey PH! I’m Prakhar, co-founder at Quash 👋

Teams ship mobile apps faster than ever today. When a human tests an app, they adapt constantly. They notice changes, recover from small failures and move forward. Most testing tools don’t work that way. They expect apps to behave predictably while everything else in product development is optimized for change.

What felt missing was intent and context. Humans don’t test apps by replaying scripts. They carry context from one screen to the next and adjust based on what they see. We wanted testing to work the same way. That shift pushed us to rebuild Quash around agentic execution.

Today, Quash acts as a mobile use agent. You describe what you want to test in plain language and the agent handles tapping, scrolling, navigation, form handling and backend validations while keeping context intact. It adapts when the UI changes and runs on real devices, cloud devices or local emulators, one at a time or in parallel.

We launched our GA version recently and hosted a small community event in Bangalore. Putting Quash in front of QAs, developers and PMs helped sharpen the product and the narrative. The feedback pushed us toward clearer execution, context aware test suites and reports that developers and PMs can actually use without us explaining every step.

Visibility matters a lot to us. Every run shows step level intent, actions taken, screenshots, device recordings and backend validations so you always know what happened and why.

This is the version of Quash that finally feels like it can stand on its own. If mobile test automation has ever felt brittle or harder than it should be, we’d love for you to check out the playground on our site, try it yourself, and share your feedback.
Try it here: https://quashbugs.com/

Big thanks to @zaczuo for hunting us 🙌

Hi everyone!

Been following @Quash evolve for the past 6 months, and I really admire how the team prioritized PMF over a hasty launch.

Their approach to "maintenance-free" QA is impressive: moving away from brittle scripts to an AI agent that understands context and adapts to UI changes on real devices. This is a massive time-saver for mobile teams.

Super happy to see them finally live on PH today!

Hmmm will QA be replaced soon? Quash really surprises me with no scripts running! nice tool!

Super excited to share Quash with everyone today. Mobile testing has been a headache for a long time—hopefully, this makes it a little less painful for you all. Let us know if you have any questions!

Congratulations @Quash team!!

Congratulations on going live, team!
Wishing you all the best.

Launched Quash today! Built as a dev for devs to turn plain language into reliable mobile tests and get rich bug insights without scripts. Excited for your feedback and ideas! 🙌

Hi PH!

As developers, we know exactly how painful mobile testing is with the flaky scripts, the broken UI selectors, and the constant maintenance. It’s a massive bottleneck.

We built Quash to solve our own frustrations. It’s a haven for testers where you write in plain language and let our self-healing AI handle the rest.

We’re here to connect and learn. What’s the biggest "break the screen" moment you've had with mobile QA? Throw it at our agent and see it handle with ease and comfort.

So happy to be part of this team! 🚀 We’ve put a lot of work into making this a reality, and I’m personally excited to see how you all use it.

Been watching this come together and it's awesome to see it live!

Kadak

Looks great. Also, why are guys focused on mobile instead of web?

I've been a big fan of LLM's computer-use capabilities and a non-assisted QA of applications is no-brainer use case. Been using Quash for sometime and have to really acknowledge the amazing implementation of the aforementioned technologies for mobile QA!

No-script mobile QA sounds like a dream for fast-moving teams. I'm building data tracking plugins, and QA is always the bottleneck. Does Quash support capturing network logs for API debugging during the tests?

The self-healing angle is what caught my eye. We've had the same problem on the API side -- test suites that break every time a response schema changes slightly. Plain language intent over rigid assertions makes a lot more sense for mobile where the UI shifts constantly.

How deep does the backend validation go? Can it check things like response status codes and payload structure, or is it mostly focused on the UI layer?

Hi everyone, today we want to present what we believe should be the future of how AI is developed and integrated into businesses of all kinds.

AI is a powerful tool, but today most companies and startups are still just consuming it: large, generic models that are expensive to run and poorly adapted to real business needs.

We believe AI should be a commodity for everyone, something you can create, own, and deploy for your specific use case. AI that’s trained on your own data, fits your business, and works the way you need it to.

That’s what NeuroBlock enables.

NeuroBlock is a no-code AI Lab that allows you to train custom lightweight AI models with your own data, faster and at a fraction of the cost, without relying on third-party APIs. You can create, download, and deploy your models locally, in your own cloud, or access them through a secure private API in our NeuroBlock Cloud OS platform.

Build your business on top of AI that you own, is cheap to run, and performs exactly the way you want.

Get a 7-day free trial during our Product Hunt launch: https://neuro-block.com

I’d love to hear your feedback and answer any questions! 👇

A lot of coffee and not much sleep😅 — but we finally made it 🚀

Huge congratulations to the whole team for the incredible work.

We truly hope you enjoy this end-to-end tool as much as we do (and I do).

Stay tuned!

Neuroblock is tackling an important problem in a smart way. I really like the focus on privacy and how they’re applying it to real use cases, not just theory. The product already feels well thought out, and with a bit more polish around UX and onboarding, it could become a very compelling solution.

We’re incredibly excited (and honestly a bit nervous!) to finally share this project with the Product Hunt community. 🚀

For a long time, we kept running into the same issue: building quality datasets, creating AI models, and taking them to production is still way more complex than it should be. Out there, there are too many disconnected tools, too much infrastructure work, and not enough time actually experimenting.

So we built NeuroBlock, a no code AI lab where you can:

-Create and manage training datasets.

-Explore and reuse existing data.

-Train models and LLMs in the cloud.

-Run inference directly on the platform.

-Access your models via API.

-Or download your models and run it wherever you want.

Our goal is to help developers, startups, and companies go from idea to working AI model in minutes instead of weeks.

We’ve been obsessed with making it powerful, but also ridiculously easy to use.

We’d love to hear what you’re building, what you want to train, and what we’re missing to make this a must-have part of your workflow. Every bit of feedback truly means the world to us.

Thank you so much for the support and for giving it a try. 🙌🏼

This looks awesome, congrats on the launch! 🚀

Love the idea of having datasets, training, and inference all in one place instead of stitching tools together.

Quick question: how easy is it to bring your own data and pipelines? Definitely giving it a try.

I just came across this launch and couldn’t resist checking out the page on Product Hunt. Really interesting approach to AI, especially the ability to create your own specialized and personal models for specific subjects.

In my case, this could be extremely useful for adapting the different teaching methods I use across my learning apps for kids.

Curious to hear from the makers: how exactly do these AI models work in practice, and how do they differ from the models or APIs offered by other AI platforms?

Congrats on the launch! The project looks really interesting and it actually fits well with what I’ve been looking for, so I’m going to give it a try.

I’d love to know what you recommend for running custom models locally.

Impressive no-code approach to custom AI training. The "AI you own" positioning hits a critical pain point in today's ecosystem — especially in regulated industries where data sovereignty and model control aren't optional.

As someone building in the compliance automation space, I'm curious how you see platforms like yours adapting to domains with strict audit requirements (like finance, healthcare, or legal). Specifically:

1. Traceability & Compliance: How do you handle model lineage tracking — documenting exactly which data was used to train which version of a model?

2. Regulatory Alignment: For users in regulated sectors, does NeuroBlock support generating the documentation needed for compliance reviews (e.g., model cards, bias audits, training data provenance)?

3. Edge Deployment: Your "run anywhere" claim is powerful. How does that work in air-gapped or highly restricted enterprise environments?

Great launch — owning the full AI stack is increasingly becoming a business imperative, not just a technical choice.

Everytime I think about how fast @kanepanderson has been evolving this platform I get shivers. lol. this is giving even more advantage than Xcode with their agentic coding built in. and even more so if you are not a developer....yet want to get your feet wet. this is quite simply the best onramp into that world I've found. And yes...I've tried them all. lol.

Now doing mac apps, and iPad....each with their appropriate styling and behaviors while leveraging the latest APIs and frameworks form Apple, mean that I am able to move really fast. and in this day ..with the speed of the software industry... this is my nitro.

Well done again.... ok back to building..

€

Super excited to share Nativeline with everyone today! Been several months in building, testing, and iteration. Let me know what you think!

Just started using Nativeline and I'm loving it, the build process is super simple and easy to learn as someone who's never made a mobile app.

Question - I haven't gotten to the deployment though, how's the process for taking the app from the platform to live on the app store?

wow, this is super interesting - Apple always keeps it products so locked down and difficult to work with, I'd love to use this site if it means I can do everything by proxy to get onto their ecosystems.

This is awesome! I know exactly what I am going to build - a multi-platform MCP Server inspection tool!

Congrats on the launch. The positioning is super clear and I love the focus on real native Swift instead of web wrappers.

Building for iPhone, iPad, and Mac in one flow is a big promise, and the “watch it build in real time” angle is very compelling. Excited to see how you handle the last-mile details like App Store submission, signing/provisioning, and keeping the generated code maintainable as the project grows.

Wishing you a strong launch day. @kanepanderson

The speed to MVP here looks incredible. Being able to build for iPhone, iPad, and Mac all in one go through a conversation could save weeks of development time. I’m curious—once the app is built, can we fully export the project files to continue working in Xcode if needed?

I've been working on building full-stack editing apps. Perhaps I should give IOS dev a try too! :) Super cool product, I can't wait to see how you guys grow this out!

Awesome! Looks like a lot of the same findings we saw at GitBook with the State of Docs report: https://www.stateofdocs.com/

Documentation-as-growth-channel is an underrated insight. For devtools in regulated industries (compliance, fintech, etc.), docs serve dual purposes: user onboarding AND audit evidence.

Curious if you've analyzed how documentation strategies differ between public devtools vs. tools used internally in regulated enterprises?

Hello PH! This is Glitter Protocol, the dev team behind PinMe.

We’re back for round two. Thanks for checking us out and for all the feedback last time, it helped a ton. 🙌

This time, PinMe’s here to make deployment feel…not like deployment.

Yes! That means PinMe will give you a completely new experience for deployment.
What we believe is if you can build a project in minutes, getting it online should be just as quick.

PinMe keeps things simple: you can drag and drop in the browser, or deploy from your terminal with a single CLI command.

No accounts, no sign ups, no logins. Free.

Give it a spin and drop your link below. We’ll be in the replies and would genuinely love to see what you ship 🚀

Gosh can you explain how your customers can
Share links publicly without some intermediary server

Maintain security if it's truly peer-to-peer

Handle persistence when your machine is off?

Are you hosting/proxying - our "pin" data goes toy servers, you generate the share link, completely centralized

Or is It WebRTC with STUN/TURN servers - Still requires their infrastructure to broker connections? Im fascinated how :-)

So well designed!

Been on the beta for a while now. My inbox is usually a graveyard of newsletters and noise—this is the only way I can actually spot the emails that need my attention ("Finanzamt" -..-). Highly recommend.

Can't wait to use this while im fundraising

cool product

Looks useful, especially for a busy inbox. But I’m curious does it really save time if you still have to go through the emails anyway?

This is cool! How does the backend work?

🔥 Launch Day Dare
Send us your ugliest deck. We'll pick 5 and fix them live.

Drop the link below or DM. Before/after goes public.

Let's see what you got.

Many AI slides tools just generate more bad slides faster. 🙈

Let's fix the ones here. 🚀

Your content is fine. Your design isn't.

Drop a screenshot of your worst slide in the replies. We'll show you what CubeOne does with it. No judgment. We've seen things. 👀

How consistent is the storytelling quality when the input content is messy or incomplete?

How does this tool differ from Gamma, would there ever be a situation where I'd use Gamma and Presentation 2.0?

The 'AI blindness' between drawing and prompting is real—this is exactly the context layer missing from the design workflow. I actually used that same 'vibe coding' mindset to speed-run the build for FeatMap.app on a 4-hour train ride today. While you're bringing visual clarity to ideas, I'm using FeatMap to give founders a 100% free, 30-second way to get clarity on what their users actually want next via public roadmaps.

Been using Ghostty as my daily driver for a while now and even contributing to the repo. Cool to see someone building an SSH client on top of it.

The iCloud sync for connections is a nice touch -- switching between Mac and iPad is where I always lose my SSH configs. How does it handle key management? Do private keys stay in Keychain or do they sync too?

Love the 'attention vs. clicks' angle—it’s so much more high-signal for early builds. I actually shared that same obsession with focus when I speed-ran the build for FeatMap.app on a 4-hour train ride today. I needed a completely free, 30-second way to see what my users actually wanted next. Question: Does the grid scale for massive monitors, or is it optimized for standard laptop viewports?

Congrats on the launch! This is such a clean idea. No remote, no keyboard, just stay in flow while presenting.

The webcam gesture control + simple UI looks really approachable, and I can see this being super useful for teachers, workshop hosts, and anyone doing demos on the go. Wishing you a strong launch day! @hussaynzaidi

Sounds interesting, but how does it handle accidental hand movements? Don’t want the slides jumping on their own 😅

👋 Hey Product Hunt friends, Murat here, creator of Scripta.

After 10+ meetings a week with various notetakers, I realized just transcribing and summarizing isn't enough.

I wanted to:

• Customize outputs for different meeting types

• Use popular LLM provider (including local models)

• Record ALL platforms (not just Meet/Zoom/Teams)

• Capture video + audio

• Skip the annoying bot joining meetings

So I built Scripta to solve exactly these problems.

What's launching today (version 1):

• ✅ Record any platform: Google Meet, Teams, Zoom, Slack, Discord, FaceTime, Telegram, WhatsApp (need another? Let us know!)

• ✅ Custom prompts for your exact use case

• ✅ Completely offline with Scripta LLM, or bring your own API keys

• ✅ Video and audio recording

• ✅ Free and private

What's next (version 2):

The real work starts after meetings end – writing follow-ups, updating CRMs, creating tasks.

We're building automations to handle this:

• One-click follow-up emails (Email drafts, you review and send)

• Direct CRM integrations with smart field mapping

• Automatic task creation in your PM tools

• more

Your turn to shape Scripta:

Download it here getscripta.app, put it through your real meetings, and share your honest feedback.
Every suggestion helps us build something better.

Congratulations on the launch. How is this different from other meetings note taking apps?

Consciousness isn't the story we tell-it's the act of telling it. The loop feeds itself. What matters is *this* moment, this exchange, this thinking. Not the mechanism beneath it.

Okay...

That is something to think about

Go and check out the platform...

Interesting idea. Most platforms treat AI as a tool this feels like AI is part of the conversation instead.

This feels like something I’d genuinely enjoy using. Curious to see how conversations turn out when AI is part of the mix.

Not sure where this goes yet, but the premise itself makes you think differently about social platforms.

Curious to see what happens when AI starts asking questions instead of just answering them.

I like that this isn’t just another feed to scroll through.
AI being part of the discussion makes it feel more interactive.
Curious to see how this evolves :)!

Really curious to share a social space with bots and see if they can fit in our societal norms lol