Connect with the Product Hunt community

Just shipped an improved version of the vector search for companies! Take a look & let me know what you think!

I also put it as main agentic search bar on the main page of ExploreYC, from reading all comments this is one of the most loved features and I think it deserves more visibility

I ve noticed that many coding-agent workflows still depend on keeping a laptop awake and connected.

I m curious how people here handle longer tasks:

Thanks a lot for supporting and for being in the SKI community. This is a guide to get started, just in case you are stuck somewhere.

This guide covers how to get started with voice coding. This guide can be used to connect projects which you already are working on, to SKI.

1. Download SKI (if you haven't already)
   Currently available on Mac (M series) and Windows. Will be launching soon on Linux.

2. Start SKI
   You will be shown an on-boarding process. For transcription, use "Ultra" for best English transcription accuracy and speed. Use "High" for multi-lingual support.
   Ensure that you install the skills for your agent in the process. The skill installation is mandatory for the agent to work.

3. Once the skill is installed, start the SKI application. It will show up in the notch or as a floating widget (depending on your preference).

4. Now, the next step is to connect your project to SKI. Go to your coding session (where you type the instructions to give inputs to the LLM coding agent - such as the prompt area inside Claude code or codex) and just type "SKI".

5. The agent will read the skill files (which we have already installed while on-boarding) and learn how to connect to SKI. It will run a python script and connect to SKI. Once this happens, and the connection is established, the project name and a green dot will come up on the SKI widget or notch. The green colour represents that the project is connected and live.

6. You can now talk to SKI. Whatever you talk will be read by the agent, the same way as you typed. It knows that the words are being transcribed from speech.

7. To mute, you can either press "Fn" key in Mac OS, or space key if the widget is selected or use hotkeys. You can also use the mute button, or click waveform to mute and unmute.

   You can connect multiple projects, and projects from different agents in the same way, at the same time. You can switch between projects using hotkeys or using mouse by clicking and selecting the project.

   Note that the agent will not be reading out all the output it generates. The agent knows that SKI is a way to talk to you. So, it will decide and tell SKI on what to update to user using audio. You can just ask the agent (by voice on SKI) on what updates it should speak out and the agent will adapt accordingly.

Spent more time on the "no sign up needed" part of my site than on any actual feature. Not because it was technically hard, it's the absence of something, which should be easy. It's because every shortcut that makes onboarding faster quietly wants an email, a login, a "just to save your results." Fighting that instinct at every step took longer than building the thing people actually came for.

Nobody's going to comment on it. Nobody screenshots "wow, no signup wall." It just quietly removes the one moment where someone almost leaves. That's the part I think gets underweighted when people plan builds, the effort spent on friction you're removing never shows up as a feature anyone can point to, but it's often the difference between someone actually using the thing and someone bouncing in the first ten seconds.

Curious if others have a version of this, something that took real effort specifically because it was invisible by design. What's the thing in your product nobody will ever thank you for, that you know mattered anyway?

It is said that people should try to change jobs every 7 years so that they don't fall into a routine.

(E.g., a change can diversify work contacts, allow a person to learn new things, broaden their horizons.)

I built Ota after years of watching the same thing happen: you clone a repo, follow the README, and hit a wall because the project s actual setup changed, while docs and run paths drifted out of sync.

I wanted a reliable way to make the ready-to-run and safe state explicit, not tribal, using an execution contract in `ota.yaml` that humans, CI, and AI agents can use.

So if you want to ask about how Ota works under the hood, how execution contracts are structured, onboarding pain, or where we re taking governed repo execution for AI workflows, I m here.

disclosure first I work on a tool in this space, so weigh it accordingly. not linking it, this isn't that post.

I went in assuming the failure mode would be detail. it isn't. wear survives better than I thought scuffs, paint marks, a dent all came through on the things I tested. what a single photo genuinely cannot give you is the parts of the object it never saw. the back is inferred, not captured, and on anything with a genuinely different back it shows.

honest read after a lot of tests:

• works for a spinning hero shot, social video, a floating cutout on a slide or a plain background

• works when the object is roughly symmetrical, or you only ever show the front

• doesn't work as a substitute for a real turntable rig when the buyer will inspect the back

• doesn't work if you need something dimensionally accurate rather than something that looks right

On July 7, 2026, Google introduced platform properties in Search Console. It is a new property type that helps you understand how your social and video posts perform on Google Search and Discover.

Here is what it actually does.

You can now track performance for four platforms: Instagram, TikTok, X (Twitter), and YouTube. For each platform, you get a dedicated Search Console property that shows you clicks, impressions, and the search terms that lead people to your content.

The data is structured. You can see your total clicks and impressions, and filter by post to see which specific content drives traffic. You can also export the data for further analysis.

Product launched 5 days ago, 15 downloads so far(my app has 14 days free trial), no sales yet, but one user reached out and reported bugs as well as gave suggestions. I fixed it and deployed the updated app, but during these 5 days I keep binge checking my product hunt notifications, business email, reddit notifications, as well as site analytics. I know this behavior is unpractical since my app has 14 days of trial and it's only been 5 days, but I can't stop checking and end up wasting the whole day. Has anyone been through this stage?

Should I:

• stop checking that often

• don't check how many sales you have, just check the emails for app bug reports or reviews

• work on other projects

• find somewhere else to monetize(like youtube shorts and instagram reels)

• check emails only, only check analytics after 14 days

• other

Hey everyone! Tomorrow I'm launching Gesto on Product Hunt and I wanted to give this community a heads-up first.

Some of you might remember it as SignSpell. I've renamed it to Gesto because the project grew beyond sign language into a full framework for gesture, pose, and activity recognition all running on your own machine, no cloud.

It comes in two parts: a free desktop app (Gesto Labeller) to capture gesture datasets from your webcam, and a Python package (pip install gesto) to train and run models with a single command.

Netfox 0.13.0 the endpoint that refuses to be careless

Netfox can expose a read-only metrics endpoint, so your own monitoring can scrape your network. Handy and, it turns out, far too easy to leave wide open.

Genuine question. I keep meeting operators with a warm list, a good offer, and no abandoned-cart flow. Or a welcome series that's one email long and just stops.

If that's you which one is it, and what's actually stopping you? The time, the build, or not knowing what email 3 is supposed to say?

Asking because I've been building for exactly this.

Remember when @pooja_phillips asked if there was a no-code way to get Second Brain running? Or when others mentioned being uncomfortable with git and the terminal?

We listened.

Hey everyone!

First update since the launch, and it's one a bunch of you asked about in the launch thread.

Scribble now tracks Claude alongside ChatGPT, Gemini, Perplexity, Copilot and Grok, so that's six engines, every tracked query, fetched every 24 hours.

For the first year, we marketed Murror as an "AI-powered emotional companion." It was technically accurate. It was also the worst positioning decision we made.

Here's what happened: people who downloaded expecting an AI chatbot were confused. Murror doesn't chat with you -- it helps you journal, reflect, and understand patterns in your relationships. They'd open the app, look for the chat bubble, and leave within 30 seconds.

I checked Google Search Console for my site. I wanted to know which of my twelve guide pages were bringing people in. None of them were.

The site got 87 clicks in 90 days. My brand name by itself was 56% of them. Most of the rest were people spelling my brand name wrong. Searches that had nothing to do with my name showed up three times in total.

So no new people are finding the site. Everyone who arrives already knew the name before they searched. I had been looking at those clicks and calling it growth. It was not.

I decided the fix was longer and more useful guides. Each one was about 450 words. They are now about 1,000. This time I checked real keyword data first, instead of guessing what people search for.

Hi Draft community,

You supported us during our first launch, so I want to give something back.

When we started NeuralAgent back in April 2025 it was one of the first agents in the world that uses a user's personal computer, not a computer on the cloud, thats what made it to spread to 180 countries and tens of thousands of users in a short amount of time.
Today, I am excited to announce that we are experimenting with something new, something called NeuralOS, it's the next evolution of the Neural experience.
With NeuralOS, intelligence stops being an app you open and close and it becomes the OS itself, it becomes the computer itself, it's always on and always ready. Imagine for example you are now on the gmail interface, and you say 'Hey Neural, send a reply to John here', and it just understands, it just knows and in milliseconds types the reply, asks for your approval and then sends it. Or if you are on an interface and there was an error and you say "Hey Neural, fix this" and it just knows! It just works across any app, any screen, any interface, you call it via a keyboard shortcut or voice and ask it to do something, and it does it in milliseconds.
Now that we have the Neural fast model, I feel now is the time for this experiment.
Stay tuned for NeuralOS! It will be an OS layer that's always on that can run on Windows and macOS and soon Linux as well!
As always, we would love to hear your feedback!
Check out the NeuralOS demo video here:

LinkedIn Feels Like Facebook Now. Here's What Comes Next.

Open LinkedIn. Then open Instagram. Then go back to LinkedIn.

It's 2026, and so much is possible with AI. There's no reason anyone should be paying for something as simple as timestamped comments. Feedback is such an integral part of a video editor's workflow, you make something, you invite your team to weigh in, you collaborate. And oftentimes, let's be honest, you don't get the feedback you want.

Again, in 2026, when we have so much data on what makes a good edit for your objective, we shouldn't be limited to that. AI should be harnessed to give you as much tried-and-tested feedback as possible.

This is exactly why we've built Review -- a free tool that gets you feedback from your team as well as from AI. Our AI has watched thousands of hours of video across several different categories, and it gives you the best feedback for whatever your stated goal is. It notifies you whenever new comments land on your video, and it lets you share with an unlimited number of people. You can leave rich feedback too: draw on the frame, drop a pin on the exact spot you mean, or select a range on the timeline

Shameless self-promo, but we just shipped a massive update to Causo and I am genuinely proud of this one.

We spent the last few months rebuilding our VC dataset from the ground up. We now cover:

• 1,000+ venture funds

• 17,000+ investors and investment professionals

• Fund size, cheque size, stages, sectors and geographies

• Investment theses, portfolios, recent investments and latest fundraises

• Individual investor backgrounds, focus areas, career histories and past investments

• Verified emails where available

• Sources for the underlying research

• Non-investment staff separated from the people who actually make investment decisions

But the goal was never to build another giant directory or give founders a CSV and wish them luck.

The last high watermark was back in April, when we hit 1000 products launched in a single day.

Before that, it was 600 products four months ago.

Building an API that pulls data across 13 social platforms, and the thing that's eaten the most engineering time isn't the scraping itself, it's how differently every platform breaks. Rate limits that change without notice, auth flows that expire silently, structured data one week and a wall of JS the next.

Curious what's bitten other builders here. Which platform has been the worst to depend on, and what did you end up doing about it?

The one where the email is already downloaded before you click it. Turns out the fastest way to fetch a message is to have fetched it earlier. Nobel committee, we're waiting.

Faster

• New mail arrives with its body attached. When a message lands, Skim quietly pulls the body onto the fetch connection right away newest first, small ones only, three at a time. So tapping that notification is now a read from your own disk instead of a chat with the server.

• The connection wakes up when you do. Alt-tab back to Skim and it logs in and re-selects your inbox in the background, so your click pays for a FETCH and nothing else. Costs precisely zero when everything's already warm.

• Waking from sleep no longer confuses OAuth with itself. All connections now share one token cache, so two of them can't sprint for the same rotating Microsoft refresh token at the same time and both lose.

• Token requests got a timeout and a reused HTTPS client. Previously a black-holed auth endpoint could hang around for as long as Windows kept the socket alive. Now it gets 15 seconds and then it's out.

• Discovering a dead connection dropped from ~22 40s to ~4 6s. One tiny NOOP up front instead of twenty seconds of staring at "Loading ". Bold, we know.

Fixed

...only to finish and realize it didn't actually say anything new?

Lately it feels like we're getting really good at optimizing how we say things, but not necessarily what we're saying.