PH热榜 | 2026-08-03

Hey Product Hunt! 👋

I built AgentSky because production-ready AI agents require far more infrastructure than most people expect. The idea came from building tycoon.us, where we ran into the same challenges: testing multiple harness/LLM combinations (let alone harness version updates 😱), fast & secure sandboxes, surviving restarts, and connecting them to every channel users expect.

AgentSky does that plumbing for you. Pick a harness — Claude Code, Codex, Hermes, or OpenClaw — pick a model, and launch in one click, or a CLI command. Your agent runs always-on in its own cloud sandbox with full history, artifacts persistence, state snapshots, backup and restore.

It's reachable wherever your users already are — WhatsApp, iMessage, Telegram, Slack, web, CLI, or our developer API's. Same protocol, same memory, every harness, every channel.

Already battle-tested in production, AgentSky powers tycoon.us, where it has handled over 10K+ agent sessions.

Parking an agent is free — you only pay when it's actively working.

Would love your feedback, especially on which runtimes and channels you'd want next!

Suspend and resume is the interesting part, and I think it has a blind spot worth designing for now rather than later.

If parking is free and I only pay while the agent is working, then an agent that has quietly stopped working costs me nothing. Which is lovely, right up until a dead agent and a cheap month look identical on the invoice. For long horizon work the bill has historically been the thing that told people something had gone wrong, and this pricing removes that signal on purpose.

From outside, a suspended agent, a finished agent, and one that crashed and never resumed all present the same way. No activity, no cost, nothing in the channel. So what tells the operator which of the three they have? What I would want is for a long horizon agent to declare an expected cadence when it launches, so the platform can say this one should have woken by now and has not, instead of leaving silence to mean whatever the reader assumes it means.

Second one, and it falls out of snapshots combined with messaging channels. If you snapshot state and restore it, what happens when the snapshot was taken partway through an external side effect? An agent restored to a point just before it sent a WhatsApp message cannot tell whether that message went. Redoing it and skipping it are both wrong, and only one of those is visible to the person on the other end, who receives it twice.

Do side effects get recorded outside the snapshot, so a resumed agent knows what already left the building?

Love the pay-only-when-working model, that's a smart pricing call.
Curious how you're handling security across sandboxes when agents are running always-on like that?

This is exactly the kind of infra headache people underestimate until they hit it.
How long did it take you to get restart working reliably across different harnesses?

congratulations! can agents communicate with users across multiple channels simultaneously? for example, could an agent start on Slack and continue the same conversation on WhatsApp?

I like that AgentSky puts different models and agent tools in one place. It could save a lot of time when testing ideas, especially for people who don’t want to set up everything themselves. I’d still like to see more real examples before using it for a bigger project.

Congrats Darren! The channels part is what sells it for me. I correct my coding agent from the phone mid-run all the time, and that flow is usually held together with duct tape. Curious: when a WhatsApp message lands while the agent is mid-task, does it interrupt the run or queue until the current step finishes?

Congratulations on the launch!
I like the idea and after reading comments i like that context is shared between agents and channels as well as you don't pay if the agent is not used. I would like to try it next dayse with with some of my tasks.

Saying state restore is best effort and cannot prevent a race condition is rare on a launch day. You mentioned wanting idempotent agent replies in your own layer. Where do you think that line sits between the infra and the product on top, Darren?

Reaching a long running agent from WhatsApp or iMessage instead of yet another dashboard feels like how normal people would actually use this. Congrats Darren, the 10,000 plus sessions on tycoon.us before launching is reassuring too. What does recovery look like when an agent goes sideways, can I roll back to any earlier snapshot?

Can AgentSky automatically scale when multiple users start interacting with the same agent? how do you handle sudden spikes in agent sessions?

If you had to recommend one harness for someone building their first long-running agent, which one would you suggest, and why?

congrats! can developers choose different LLM models for the same harness? how quickly can you switch models without changing the agent configuration?

I like that agents stay available across WhatsApp, Slack, and the web. Meeting users where they already work makes a lot of sense.

what runtime or channel are you most excited to add next? are you prioritizing integrations based on community requests or technical demand?

Congrats on the launch!
What does the developer API expose compared with the CLI? can teams programmatically create, pause, resume and manage agents?

how does AgentSky handle failed or partially completed tasks? can an agent resume from its previous state instead of starting over?

congrats! can one agent use different models depending on the task it is performing? for example could it use a cheaper model for simple tasks and a stronger one for complex reasoning?

The biggest challenge I see with multi-channel agents is maintaining context. If someone asks a technical question in a community forum today and follows up with the same issue in a private chat tomorrow, how does the agent know it’s the same user and the same problem?

Is the conversation history unified at the agent level, or does each platform maintain its own separate memory?

The multi-channel consistency is the part I want to understand better. If a user starts a thread on Telegram and then messages from Slack the next day, does the agent see a single unified conversation or two separate sessions? For developer communities I manage, people switch channels constantly and that is usually where context breaks happen.

how does pricing work when an agent is parked but still maintaining its state? is there any cost for storage, snapshots or backups while it is inactive?

Congrats on the launch! 👏
How does AgentSky handle failed or partially completed tasks? can an agent resume from its previous state instead of starting over?

The cross-platform memory is the part I’m most curious about. If a user starts debugging a problem through Discord, then continues the conversation through email or a support portal later, does the agent maintain the same context and history, or does each channel create an isolated session?

For developer communities, users rarely stay on one platform, so I’m wondering how agents handle context continuity when conversations move across different channels.

I didn't expect the AI asset generator to handle organic shapes this well. My plant models actually look like plants.

Vinay here, founder & CEO of Ctruh

I started this company because I kept watching brilliant products get sold through flat photos that did them no justice. We've spent years on a hard technical bet: that immersive 3D and AR shouldn't require a game engine, a specialist team, or weeks of work. It should be as simple as sharing a link.

Studio is the clearest expression yet of that belief. Seeing it in the hands of the community today genuinely means a lot to our whole team.

I'll be here throughout the day, and I'd love to hear what you think, especially the hard questions. Thank you for spending a moment with what we've built.

What stood out to me was the material editor. I run a ceramics shop and getting realistic glaze textures on 3D models used to require outside software. Now I adjust roughness and reflectivity directly in the platform and the results are close enough to real photos that customers stopped asking for extra images.

Excited to finally launch this after months of building and refining it with the team.

One thing we kept hearing from brands was that they wanted 3D and AR experiences, but the workflow was too fragmented: asset creation, optimization, hosting, web integration, and customization all lived in different tools.

Ctruh Studio was built to simplify that entire process in one browser-native platform. The goal is to make immersive commerce practical, not just possible.

Would genuinely love feedback from the Product Hunt community on what workflows we should improve next.

No-code is the right call here. I've spent weekends fighting with Blender just to get a basic product spin working. Being able to build that same thing in a browser changes my whole workflow.

What caught my eye is the AR piece. I run a small furniture shop and letting customers preview pieces in their own room could cut returns significantly.

Anyone found a good workflow for batch editing multiple product variants at once? Doing them one by one is slow.

The product is fabulous! Any brand looking to create 3D assets, or launch a virtual storefront or offer Virtual Try-ons, Ctruh Studio makes it extremely easy to do that.

Even a non-techie like me was able to create a try-on in a matter for few minutes. If you have not already, try it out and have your mind blown!

One thing kept coming up in almost every conversation: We love immersive experiences, but they're just too difficult to build.

That became our starting point.

So we built @Ctruh Studio: a platform that lets anyone create and publish interactive 3D experiences for the web without the usual complexity.

We're just getting started, and I'd genuinely love your feedback.

If you get a chance to try it, let us know what you'd build first. Every comment helps us make the product better. 🚀

It's amazing to see something we've spent years building finally out in the world.

From the very beginning, our goal was simple: make immersive commerce easy to create and even easier to experience without downloads, plugins, or unnecessary complexity.

If you're visiting us on Product Hunt today, spend a few minutes exploring the live demos. The best way to understand the product is to experience it yourself. We'd love your feedback and support!

I'm curious how it handles file size optimization for mobile users. Rich 3D experiences often load slowly on phones. If you've solved that, it's a real edge over competitors.

Watching Ctruh Studio go from an idea to a product that is now live on Product Hunt has been an incredible journey. Having worked alongside the team, I've seen the countless hours spent solving complex challenges to make 3D and XR creation simple, accessible, and browser-native.

What excites me most is that we're not just launching another tool. We're enabling creators, brands, and businesses to build immersive experiences without needing specialized technical expertise. Proud to be part of a team that's pushing the boundaries of what's possible in digital experiences. Congratulations to everyone who made this happen. 🚀

The spec I'd put on the page is mesh weight, not render quality. A configurator that takes 8 seconds on mobile 4G converts worse than the flat photo it replaced, so the number that matters is poly count and texture size after your optimizer runs, not what VersaAI generates before it. If you're already holding a per-product budget, say it loudly. That's the line that decides whether a brand ships this or shelves it after the pilot.

Watching Ctruh Studio go from an idea to a product that's now live on Product Hunt has been an incredible journey.

It's amazing to see years of hard work finally out in the world. The goal has always been simple: make immersive 3D, XR and AI experiences easy to create and experience right in the browser, without downloads or unnecessary complexity.

Proud to be part of a team that's making immersive technology more accessible for creators, brands, and businesses.

If you're visiting us on Product Hunt today, take a few minutes to explore the live demos. We'd love your feedback and support! 🚀

Big congrats to the whole team

Really love how you are making immersive shopping far easier for everyone

Wishing you a huge launch day and plenty of momentum

What has been the biggest surprise from early users so far

Just tried it out, and I was genuinely impressed! I converted almost every item I've bought online just by clicking a photo, and it turned them into 3D models so quickly. The experience was incredibly smooth. I could zoom in really close, rotate the models from every angle, and explore every little detail.

Whenever I wanted to make a change or try something different, the AI chatbot was there to help. The AR feature was just as impressive. I could place each item on my desk and instantly see how it would look in my space.

I honestly can't wait to see where this product goes from here. Huge credit to the CTRUH team for building something this exciting!

The platform feels simple to use, even without much technical knowledge, and the 3D, AR, and AI features save a lot of time. There is still some room for improvement, but the overall experience has been smooth and genuinely useful for presenting products in a more engaging way.

Really impressed with the experience so far. The no-code 3D workflow is smooth, the editor feels polished, and it's great to see a product making immersive content creation more accessible. Congrats on the launch!

Ctruh is building the infrastructure for the immersive web.

We make it easier for businesses to create and deploy interactive 3D, AR, and XR experiences — from virtual try-ons and 3D product configurators to virtual showrooms and immersive walkthroughs.

With browser-native 3D technology and AI-powered creation tools, Ctruh is making immersive experiences more accessible, scalable, and easier to experience without apps or complicated hardware.

We’re helping brands turn products into experiences.

Ctruh Studio makes creating interactive 3D product experiences incredibly simple with AI. I am impressed by how seamlessly it transforms product visuals into immersive experiences, it's a powerful no-code platform that helps brands stand out and drive engagement.

Ctruh is genuinely solving a real e-commerce pain point by letting shoppers see products in 3D/AR before buying and doing it without forcing brands into clunky app downloads or dev-heavy integrations, team has built something people actually want, not just a flashy demo.

I personally like there Versa AI outputs

The 2D to 3D AI blew my mind. I went from a 2D image to a full fledged XR experience in just a couple of minutes. Superb!

What stood out to me is how accessible this makes 3D and XR for people without a technical background. Using AI to generate a model from a prompt and previewing it directly in the browser feels like a genuinely useful workflow, especially for smaller brands wanting to showcase products in a more interactive way.

@Ctruh Studio is an impressive platform that enables brands to create interactive 3D, AR, virtual storefronts, and immersive product experiences with ease. I’m proud to have contributed my time and effort to building this product at Ctruh, and it’s exciting to see how it is helping brands showcase their products in a more engaging and innovative. Its powerful all in one browser based workflow is a game changer for brands looking to deliver immersive customer experiences.

Ctruh Studio is great to make product pages immersive with the AI models that can generate 3D assets from images and also create multiple variations using the tools available in the Studio, being able to make AI content as well on the same platform where all the product assets lie makes it very easy to use.

Ctruh Studio looks like an exciting platform at the intersection of 3D, AI, and XR. Making immersive experience creation more accessible could unlock huge possibilities for brands, creators, and businesses. Excited to see how Ctruh Studio evolves.

I've integrated a few 3D, AI, and XR experiences built with Ctruh Studio into my website, and the impact has been impressive. Visitors are actually exploring the experiences instead of just scrolling past, which has noticeably increased engagement and time on site. That extra interaction has translated into better conversions and ultimately more sales. It's refreshing to see a platform that makes it so easy to create immersive experiences that deliver real business value. Excited to see where Ctruh Studio goes next! 🚀

Running Google Ads is difficult. The interface is confusing, proper optimization requires special expertise, and it is very easy to waste thousands of dollars on campaigns that result in junk clicks instead of real conversions. Unless you’re a PPC expert, the cards are stacked against you.

We wanted to make Google Ads less frustrating for small teams and solo marketers. So we built native Google Ads automation directly into Airtop. Connect your Google Ads account, and Airtop will:

👉Research keywords
👉Build themed campaigns based on best practices

👉Audit and optimize spend

👉Create automatic scheduled reports

👉Send updates directly to Slack, email, and more

Airtop works like a PPC consultant would. It understands context, connects to your CRM, and makes changes directly in Google Ads once you approve them. Any change that can affect spend requires your explicit approval first. 

We’ve been running Google Ads with Airtop since May, and it’s suggested many optimizations that we otherwise would have missed. We also don’t have to spend hours buried in the Google Ads interface anymore. We’re excited to share this tool with you!

You can try Airtop’s Google Ads integration out here: airtop.ai/google-ads.

Use promo code GAPH29 for a free month of Airtop’s Starter plan.

We’re here all day. Ask us anything!

This feels like a practical use of AI rather than AI for the sake of AI. Turning Google Ads management into a conversation could save marketers hours every week, especially when launching or optimizing campaigns. Excited to see how it performs with real-world accounts. Congrats on the launch!

I run a small e-commerce shop and Google Ads has always felt like a black box to me. I set a budget, watch it disappear and hope something works. What I like here is the "no expertise required" angle, that's usually marketing fluff but if it genuinely walks a beginner through campaign structure, that's the real barrier for people like me.

the crm connection is the interesting bit tbh. an optimizer is only ever as good as the event you point it at, and signups or checkout starts will happily train the bidding to buy more of the clicks that refund later. does a cancellation weeks down the line flow back and devalue the click that caused it, or is a conversion final once it fires?

Congrats @amir_ashkenazi and team. The hard part is often not understanding the task but actually interacting with websites and completing the steps. How Airtop handles websites that change their layouts frequently?

Congratulations on the launch! I run Google Ads on a small budget, so reporting and spend checks are what actually eat the time. Does the agent work through the Google Ads interface or through the API, and how does it hold up when Google changes the layout? I would also like to know whether this is worth using below a certain monthly spend, or whether it is built for larger accounts.

Late to this one but the auth part is the decision I keep thinking about. Most web agents stop at the public web, and the useful data is almost always behind a login. Question from the content side: when a site changes its layout, does the agent work it out again by itself, or does the instruction need a rewrite? Our scrapers break every few weeks and that is the real cost, not the setup.

The way Airtop turns plain English into a working browser agent feels like real magic, not a gimmick. Love that it actually logs in and navigates instead of just pretending to.

Hi everyone!

Qwen3.8-Max is @Qwen3’s most capable model to date, built around coding, cowork, and long-running agent tasks.

One of the tests ran for 16 days straight. Qwen3.8-Max started with an empty repo, built a coding harness, and kept working on it as new issues, feedback, and test results came in. The final repo had 265 commits, 127 PRs, and 151 issues.

It is a 2.4T-parameter MoE with 95B active parameters and a 1M context window. It already works with the major harnesses you use.

The API is live now on QwenCloud.

Open weights arrive next week. This will be the largest open-weight release in Qwen’s history!

The fact that you keep open-sourcing these models at this scale is honestly wild. Most labs gate everything behind APIs, so seeing Qwen3 drop with full weights on GitHub feels like a genuine gift to the community.

Amazing, congrats on the launch. Do you have any apps in the set that ranked and then fell off, or that shipped a flow and lost conversion?

Hi, @jaimintf Very cool site, I could see it being useful, but can you tell me how does it differ from Mobbin?

Congrats on the launch! Quick question — where is the revenue data coming from? For a few of the apps I looked at, the numbers seem a bit off compared to what I'd expect. Curious if it's estimated from a third-party source or pulled directly from somewhere.

I wasn't able to test it without paying, even though it initially seemed like that might be an option. I wanted to see whether there were apps with maps. Searching by keyword returned several apps with "maps" in their names, but it wasn't clear whether they actually included maps in their interfaces or how important maps were to the overall experience.

I'd suggest allowing users to open at least 3–5 mockups after signing up, but before requiring payment. Personally, I might have subscribed if I had some confidence that the platform contained something useful for my needs.

Hey ProductHunt,

We built Plethora to be the next home for interactive content.

Our bet - As AI democratizes programming, programs become the consumables of tomorrow.

Our feeds have been full of videos, and our attention is being rented to big tech estates.

Plethora aims to bring back attention, and let the user consume with control and agency.

Plethora offers an alternative scrolling experience. Your feed consists of -

•⁠ ⁠classic games you miss - arcades and candy crush and piano tiles.

•⁠ ⁠games you have never seen and will enjoy enjoying.

•⁠ ⁠hundreds of puzzles to scratch your brain

•⁠ ⁠fidget toys for when overstimulated.

•⁠ ⁠interactive art - to interact with like god intended.

•⁠ ⁠interactive stories.

•⁠ ⁠simulations and zero-player games and worlds to hang out in.

•⁠ ⁠educational content / active learning - to finally help you see.

•⁠ ⁠utility like tuning your guitar.

•⁠ ⁠experiences waiting to be figured out.

Every bit on Plethora responds to you.

Early love for Plethora has people using it in-between moments, to get into flow state and to deal with overstimulation and anxiety.

Early creators see Plethora being the defining platform for the medium.

Follow us along our journey.

I have poured the entirety on my childhood on this app.
You will find creating for Plethora as a newfound joy.

This is such an amazing app, with so many interactive games

I especially liked pacman and another intuitive game where we had to unentangle knots/strings between points 🧠

very interesting concept. reminds me of some games like until dawn

I love the idea of this! i like this because when it comes to art, my favorite type is interactive. One day when i become a physical artist (3D mediums), i want my art to be interactive. Fun fact, im a dancer and an AI engineer, so hopefully that helps with the demographic, im very much a polymath, just stopping at those 2.

Oh wow! This looks so cool... just downloaded. I wonder if there's some room for us to put some of our interactive experiences onto your platform?

hey this sounds really fun! im wondering if it has any effect on wait times?

The completionist in me sees "all 151, fully local" and immediately wants to leave long agent runs going just to fill out the Pokédex lol. You've somehow built the one dev tool that makes me want MORE waiting — nicely done.

the second terminal tab angle is so weird it kind of works. caught myself actually looking forward to checking back in.

Going to give it a spin! Now the waiting while claude works is a real pain and I don't see it going away for the next few years. So loved this idea, super builder centric.

Congrats on the launch, @zamarrowski🚀

Converting idle wait time into a Pokemon game is such a clever idea. Context switching for 20-second waits usually kills focus, so keeping it right in the terminal is super smart. Love that it's open source and fully local too.

Good luck with the launch today!

The thing that makes this work is that the game only advances while Claude does, which is also the trap. You're paying people in Pokémon for waiting, so a run that spins for 20 minutes and produces nothing still feels great. I'd tie the encounter to exit code rather than elapsed time. Then a failed run costs you the catch, and you've accidentally built the only test result anyone would actually watch for.

👋 Hey Product Hunt!

I got sick and tired of paying subscriptions for everything in my life.

The latest addition? A transcription app that was charging me every month to upload my conversations to their servers.

So I built the opposite.

yapyap records and transcribes conversations entirely on your machine. Speaker identification, summaries, action items, custom analysis. All local. No account.

No bot is joining your calls, and no audio is leaving your device.

The parts I'm most stubborn about:

🔒 Nothing leaves your device.
Transcription, diarization, and the AI analysis all run on your hardware. Not “encrypted in transit.” Not “we don't train on your data”. Just: it never goes anywhere.

💸 You buy it once.
No monthly fee, no lock-in, and no archive holding your words hostage. Your recordings stay searchable, exportable, and yours. Forever.

🔍 It's a memory, not a transcript dump.
Everything you've ever recorded is searchable, and you can point custom analyses (“lenses”) at any conversation.

📱 There's a phone app too
For the conversations that don't happen at your desk. It syncs to your desktop over your network and no cloud in between.

Software used to work this way. I still believe it can.

I'd genuinely love your feedback!

Own your voice again.

The Mac-only caveat I’d want to know before recommending this to my team: when you capture from a browser window, does it pull the rendered text from the DOM or does it run the OCR pipeline on the screenshot? For Notion and Coda docs specifically, the difference shows up in whether nested blocks and toggles come out flat or preserve their structure.

Yeah Broden! Markdown era is here and you're a doing an amazing job on it. Congrats on this impressive launch

I’m building Hand Wave to make sign-language conversations more accessible with the camera already on Meta glasses (and other devices too). Let me know what you think!

running the model locally on-device is genuinely a cool move, keeps the latency low and the privacy story clean. love seeing open-source sign language work actually ship to glasses instead of staying in demo hell.

Love that you built this on an open-source local-first model, since accessibility tools like this should never depend on a paywall or a connection to work.

FSBoard is a fingerspelling dataset if I remember right, individual letters rather than full signs. so is this reading fingerspelled words letter by letter, or does it also handle actual ASL signs and grammar? those are pretty different problems and "sign language into speech" reads like it's promising the second one.

That's pretty cool, but what are the actual use cases? We use Claude Code, and on our team, up to 10 people can work on the same project simultaneously, but each person is responsible for their own part. Or is this mainly designed to work around usage limits?

Love the idea of collaborating inside the same AI coding session instead of constantly sharing context back and forth. Feels like a much more natural way to work as a team. Congrats on the launch!

the decision to keep host execution authority is the right call, the second you let the guest drive, you've built a pair programming tool, not a session amplifier. curious how attribution surfaces when one person's turn kicks off a 20-step agent chain: does each sub-action stay attributed to the human who prompted it, or does the chain appear as agent output with a single human-turn anchor?

Launch-day update: v0.4.13 is live. The host still installs once; the teammate now pastes one version-pinned npx line and lands in the exact shared room — no global install, no setup, and no background service. Both founder Macs are upgraded, and the clean Homebrew lifecycle passed on macOS and Linux. The metric that matters next is outside use: first 10 two-person teams, median invite-to-room under 5 minutes, then a later-week return. Test checklist: https://github.com/godfaddaai/multiplayer-ai/issues/7

The gestures worth shipping aren't the happy ones. Empty state, loading, and error are the three moments an app actually needs a face, because that's when it's failing the user and the mascot is the only thing buying patience. Everyone builds wave and thumbs up first, then ships a plain grey spinner anyway. And agreed with Christoph on the paywall. A mascot is a taste purchase and you can't sell taste off someone else's sample character.

Congratulations on the launch!
Does a pack keep one character consistent across all gestures, or is each gesture generated separately? Also, what does the license allow for commercial use inside a paid product?

the fully-automated pipeline is the part that gives me pause, not the AI voice or captions specifically. YouTube and TikTok have both been tightening enforcement against mass-produced, low-effort AI content this year, and "pick a topic and a schedule, we post everything" sounds like it's optimized to get a channel flagged rather than to build one that lasts. is there a human review step anywhere before something goes live, or does it genuinely post on autopilot end to end?

The risk here isn't video quality, it's that nothing in the loop has a human before publish. One Today I Learned script that states something false confidently goes out under the user's channel name and they find out in the comments. I'd make hold for review the default for the first 10 videos, then let people switch it off once they trust the scripts. Costs you nothing on the demo and saves the accounts that would have churned angry.

This is a real gap for teams, not just individuals. Half the useful context in my company now lives in one-off chats nobody can find two weeks later. Does Inventory work across a team's conversations with permissions, or is it scoped to a single person's history for now?

Got there from your comment! :) Wishing GL!

The local-only decision is the interesting one. I've got two years of Cursor and Claude Code history sitting in places I can't search, and the reason I've never tried a cloud tool for it is that those conversations contain client context I'm not willing to upload.

Two things I'd want to know: does it index historical conversations or only from install onwards? And when a tool changes its local storage format in an update, how much breaks?

Second one matters because that's the failure mode for anything indexing someone else's files, and it's the thing I'd need to trust before paying a one-time fee.

Upvoted, saw your five in Nika's thread, which is how I found this.

no cloud is a real answer to one worry but it kind of creates a smaller version of the same problem: right now my secrets and half-finished thoughts are scattered across five separate tools' storage, which is annoying but also means no single file on disk is that valuable. an index of literally every agent and IDE conversation in one place, searchable, is a much juicier single target for anything else running on the machine. is the index itself encrypted at rest, or is it plain text once it's built?

Hey Product Hunters 👋

I'm Neil, maker of Inventory.

What inspired me to build this

I kept losing conversations. Not because I deleted them. Just because they pile up across Cursor, Claude Code, Zed, Codex, Kiro... and there's no way to search any of it.

I'd vaguely remember solving something in a session from two weeks ago. Which tool was it in? Which project? No idea. So I'd spend 20 minutes hunting, or just solve it again from scratch.

Browser history doesn't help. The apps themselves have weak or zero search. And honestly I didn't want some cloud tool indexing my private dev conversations anyway.

So I built Inventory. It runs locally, indexes everything on your machine, and gives you one search bar across all of it. No account, no sync, no subscription. One time and it's yours.

The problem

If you use more than one AI tool, you know how this goes. A conversation where you figured out a gnarly bug. A prompt that finally worked. An architectural decision you reasoned through with Claude. It's all sitting somewhere on your machine but it might as well be gone.

AI agents are becoming a real part of the workflow now and nobody built the memory layer for it. That's what Inventory is.

Happy to answer anything.
Would love to hear how others are managing this 🙌

Omg, finally! I've been wishing for an easier way to edit LaTeX resumes for ages. Definitely giving this a try... Wondering if full LaTeX import/export is supported?

The live preview is instant and the markdown rendering looks clean out of the box, no setup needed. Honestly surprised how lightweight it feels compared to opening a full LaTeX editor for quick docs.

The switch won't be decided by the writing, it'll be decided by the one page you can't reproduce. Anyone submitting to a journal or a university has a template they didn't choose, so if Doxy gets 95% of the way and then can't do the numbered equations or the bibliography style, they write in Doxy and finish in LaTeX, which is worse than never switching. I'd name the document types you can fully produce today and say nothing about the rest. Shorter page, much easier decision.

Hey Product Hunters 👋
I'm Sigit, maker of MacDupl.

What inspired me to build this
I was tired of logging out of my personal Slack just to log into a client's, over and over 😭

Tried browser profiles, separate Mac users, extra browsers — all annoying in their own way.

What I actually wanted was dumb simple: another copy of the app, its own icon, own login, sitting right next to the original.

So I built it. macOS actually makes this possible with APFS copy-on-write — cloning a 700MB app costs basically nothing on disk.

A few weekends later, MacDupl was born 🚀

The problem
If you use the same app for work and personal — or juggle multiple client accounts in one tool — you know the loop: log out, log back in, log out again. Browser profiles half-solve it for browsers, but there was nothing for native Mac apps. I wanted a real second instance of the app itself: separate login, separate data, its own icon in the Dock — not a workaround.

How the approach evolved
Early on I looked at container/VM-based isolation to sandbox apps more heavily — more "bulletproof" in theory, but it added real complexity and UX tradeoffs that worked against the whole point (this was supposed to feel instant, not heavy). I ended up going with a lighter approach: unique bundle identifiers plus separate data directories per clone, using `--use-mock-keychain` so credentials stay isolated without ever touching the shared macOS Keychain. It's simpler, and it's fast — clones build in about 2 seconds.

That constraint also shaped what MacDupl *won't* do: some App Store / sandboxed apps (their entitlements and shared containers make real isolation impossible without Apple's own provisioning) just can't be cloned. Rather than fake it and quietly share the original's data, I decided to exclude them outright and be upfront about it. Sandboxed apps aren't supported — accuracy mattered more to me than a longer compatibility list.

Where it landed:

• 7-day free trial, unlimited clones

• After that, 1 clone stays free forever

• $19 one-time unlocks unlimited — no subscription

• Zero telemetry, zero analytics — the only network calls are license and update checks
  

  Would love your feedback, especially on apps not yet on my https://macdupl.app/compatibility or anything odd on the native-app tier. I'll be around all day 🙏

Two Slack accounts is the honest use case, but the interesting one is what "fully isolated" means to an app that assumed it was alone. Keychain items are shared by bundle ID, and a lot of Mac apps write to fixed paths in Application Support rather than their own container. Do you rewrite the bundle ID per clone, and if so what happens to the license — does the second instance ask to be activated again?

I love the idea. I am curious if you have thought about using this for other sports? I recently started playing tennis again and would love something that could analyze my swing and give me feedback. The setup my be a little tricky but you could have airpods in and get realtime feedback.

respect for shipping this rough and asking for harsh feedback. the part I'd worry about is camera angle on the heavier compound lifts - a phone propped up to one side can miss depth or spinal rounding on a squat or deadlift that a trainer standing at the right angle would catch immediately. if the app gives a thumbs-up on form when the angle just didn't have visibility into the thing that actually matters, that's arguably worse than no feedback at all, since people will trust it and load up more. is there anything that tells the user when the angle itself isn't good enough to judge a given movement?

congrats on the launch! I'm wondering what the different applications of this tool are or if it can analyze all types of movements.

Hey Product Hunt. I'm Moss'Ab, and I built Murmell.

Last year at a hackathon we had three people, four AI agents, and somehow still one laptop everyone was afraid to close.
Half the night went to "who's on which file" and "wait, my agent died when I shut the lid."

So I built Murmell: a shared cloud canvas in the browser where Claude Code, Codex, and your team work on one repo, on a cloud machine. Close your laptop, they keep going. Same room, a preview in the canvas, and the work lands back in git.

More is coming next, including iOS and Android simulators in the canvas so you can see mobile builds the same way.

I built most of this with Claude Code. Once it was solid enough, I started building Murmell with Murmell.

Launch offer: free for the first week, plus $60 of Claude Code credits.

Curious: when you ship with agents on a team, what breaks first for you, the coordination, the laptop, or the merge mess at the end?

murmell.com

File level claims solve the loud half. Two agents can hold completely disjoint files and still both rewrite the same contract, one edits the caller and one edits the callee, nothing conflicts and the build breaks somewhere neither of them was looking. The thing I'd want on the page is what happens to a claim when the agent holding it dies. Nobody's at the keyboard on a cloud machine, so a stale lock is silent and it's sitting on exactly the file the next person needs.

you asked what breaks first. from the shared inbox side it is coordination, but not the part people expect.

the merge mess is loud. git tells you, someone fixes it, everyone learns from it. the failure that actually costs you is two actors doing the same work without colliding at all. both finish, both succeed, nothing conflicts, and you find out downstream or not at all.

we get that version every day. two people open the same conversation, both answer, the customer gets two different replies and neither of them did anything wrong. in code you would at least get a conflict. when the work leaves the system, you just get two sends.

agents sharpen it, because a person hesitates when they see someone else typing. an agent does not hesitate and does not look.

so does the canvas have a way to claim something, a file, a task, whatever the unit is, or does it show presence and leave the rest to people noticing? presence tells you someone is in the room, not what they have taken.

Shoutout to indecisiveeric.com/gesture-synth, which went viral a few weeks ago on Instagram. It allows users to use both hands to perform chords.

gesture.live is an exploration on how to take this gesture synth further. I am using the same chord language, but am compressing it to one hand, so the other hand is free to play melodies, add drum/bass patterns and control effects.

the demo video looks great in a well lit room at a desk. has anyone actually tried this on a small stage or at a house party yet? I'm curious how the hand tracking holds up under stage lighting or just dimmer, more colored light than a webcam usually gets pointed at, since that seems like the real test of whether this leaves the bedroom-producer use case.

I am no musician, so probably I will not ever use this, other than to play with my kids, but this seems amazing!

The question that decides this is whether note-ons are quantised. Webcam hand tracking is 30fps on a good day, so there's 30ms of jitter before the audio thread even hears you, and you start feeling timing errors well under that. Snapping to a 16th grid makes it playable but you lose pushing and laying back, which is most of what makes a part sound human. Freeing the second hand by compressing chords into one is the right call either way.

Before the world tells you what matters, what deserves your attention today?

honestly the two minute watering thing feels kind of nice, like it actually forces you to slow down instead of just setting a reminder and ignoring it. my little sprout showed up after a couple days which was a small but real moment.