PH热榜 | 2026-04-04

Google Vids just leveled up video creation with major new features!

It is Google's intuitive video editing suite that solves the complexity of creating high-quality videos by letting anyone generate clips from prompts or photos , now with AI-powered video generation (Veo 3.1), custom music (Lyria 3), and fully customizable AI avatars.

What makes it stand out is the combo of free, built-in video generation plus tightly integrated tools: Lyria 3 custom music, AI avatars, a Chrome screen recorder, and direct YouTube publishing, all inside your Google ecosystem.

Key features:

• High-quality video generation from text or images using Veo 3.1, free for all Google accounts with a monthly quota.
  

• Custom music tracks with Lyria 3 / Lyria 3 Pro for Google AI Pro and Ultra subscribers, so your videos get original soundtracks that match the mood.
  

• Directable AI avatars you can place into scenes, interact with props, and fully customize in look, outfits, and backgrounds while keeping a consistent face and voice.
  

• A Google Vids Screen Recorder Chrome extension to capture your screen and camera from anywhere on the web, plus one-click publishing straight to YouTube with private-by-default uploads.
  

• Higher limits for Google AI Ultra and Workspace AI Ultra accounts, with up to 1,000 Veo videos per month for power users.

It’s ideal for creators, educators, marketers, side‑hustlers, and everyday users who want quick tutorials, social clips, birthday montages, travel vlogs, product promos, or internal explainers without learning pro editing tools; faster and cheaper than traditional editing.

Try all new features at vids.new!

It comes down to about 10 videos per month using the Veo model for free? Can a prompt also refer to video sound?

Hey everyone! 👋

The problem with most analytics tools? They're either too bloated, too expensive, or they hand your visitor data straight to Google. We built Sleek to fix that — paste one line of code and you're watching real visitors on your site live, with zero cookies and zero consent banners. Simple, fast, and actually yours.

Went through the signup and the live demo. The setup is clean, three steps, no friction and showing a real dashboard before you even create an account is the right call.

One thing I noticed going through the dashboard: it tells you who came and where from really well. What's missing is whether they're coming back. There's no return visitor rate, no new vs returning split and no scroll depth. So, a founder looking at 3 minutes average session duration has no way to tell if that means people are reading everything or just leaving a tab open. For early-stage founders trying to figure out if their messaging is actually landing, those three signals matter more than knowing 43% of visitors are from the US.

Really like the simplicity but how are you handling repeat visitor tracking without cookies?

The no-cookie, no-consent-banner approach is such a relief. I've been looking for something lightweight to replace GA on a few side projects — $5/mo forever is a great price lock too. Curious how the real-time visitor view works at scale?

The one-line-of-code setup is what sold me on trying this. I have been looking for an analytics tool that does not require a PhD in configuration just to see who is visiting my site in real time. Curious about one thing though: how does Sleek handle high-traffic spikes? Does the real-time view stay responsive when you are getting thousands of concurrent visitors?

This sounds interesting, how does this compares to something like Post hog?

Love the live view concept. Noticed "Chat with your data" in the comparison table but couldn't find details on the site. How does that work exactly? Like asking questions in plain English about your traffic?

Congratulations! The product and landing page look great. Thank you for having the live demo button to test around. One important aspect for me in a web analytics tool is the ability to export data and/or interact with it using different APIs. Is this possible with Sleek Analytics?

Really dig the approach here. One line of code and you're watching visitors live, with no cookie banners to worry about. The $5/month forever pricing is a nice touch for indie makers too. Any plans to add event tracking or conversion funnels down the road?

most analytics tools are built around retrospective summaries. you check them every morning, scroll through weekly trends, and session counts from the day before. the 'real time' view exists but it's usually secondary, showing a visitor count and recent page views with a 30-second lag, not actually watching people move through your site.

what sleek is doing differently is treating the live view as the primary interface, not an afterthought. when you just shipped something new or you're watching a landing page perform for the first time, seeing actual visitor paths as they happen is more useful than aggregates. the cookie-free fingerprint approach also means you get data from the first visitor, no consent banner delay, no waiting until compliance is sorted before you can see whether your copy is working.

How are you handling privacy and compliance if there are no cookie banners, especially with regulations getting stricter?

The no-cookie angle is huge for indie makers — one less consent banner to worry about before launch.

I'm about to ship my first iOS app and was dreading setting up analytics for the landing page.

Does Sleek work well for low-traffic sites in the early days, or does the live view feel empty until you get traction?

@uaghazade Cookieless analytics is a hard sell to people who've been on Google Analytics for years. How are you handling that objection when people land on the pricing page?

could I also utilize this for an iOS app?

What is the difference between your product and Microsoft Clarity ?

Coolest launch of the day for sure! Do you think surf feeds could become a new kind of owned media asset for creators (like newsletters or domains etc)?

I'm just back from the AtmosphereConf, where I traveled with @mmccue and Ryan Barret by overnight train to join others in work to decentralize the social web!

And now Mike and @marci_mccue are here to launch something that's been over a year in the making: Social Websites.

The idea should be familiar, but the implementation and the scale is what's new.

Now anyone can build and launch their own corner of the internet and pull in social content from @Bluesky , @Mastodon , @Pixelfed, @Threads (and more!) with familiar publishers like The Verge, Decoder, Vergecast, Version History, WIRED, Rolling Stone Politics, 404 Media, Shutdown Fullcast, The MMQB, Defector: Sports!, All Net, FilmFeed, and The Oregonian participating.

Making a Social Website

Using surf.social, anyone can make a Surf feed, publish it with a subdomain and launch it as their own social website.

Feeds can include sources from Bluesky, Mastodon, Threads, YouTube, podcast services, blogs and newsletters (RSS). Users can also assign a community hashtag, allowing people to contribute to the feed simply by using the hashtag. They can also set filters and exclude profiles or terms to keep the conversation on topic.

More tools to customize (e.g., adding custom headers and colors) and manage feeds are coming soon. Most importantly, social websites can now be shared outside of Surf, with online communities or linked in websites.

The Android app is now available and the iOS app is coming soon.

This is a big unlock for independent creators and communities. Pulling together Mastodon, Bluesky, Threads, and RSS into one curated social website is exactly what the decentralized web has been missing. Are you seeing more traction from publishers or individual creators so far?

Love the idea of giving creators a real destination they own.

If I understand correctly, you still post on Bluesky or Mastodon and Surf just organizes and surfaces that — is that right? Wondering how discoverability works for a solo indie maker just starting out vs. established brands like The Verge.

Three years ago, when we first saw Mastodon and learned more about the open protocols it was built on, we realized how powerful a shared open social graph could be, putting the power of our connections back in the hands of people. We decided to join the movement and opened ‘federated’ Flipboard, sharing our editorial curation with the fediverse. Once we had integrated our backend with the social web it was time to rethink the front end. That's what lead to Surf.

After 2.5 years of development we’re finally launching Surf, a new kind of browser for a new kind of web: the open social web. Surf lets anyone easily discover, browse and make feeds on the social web which bring together people, podcasts, videos, newsletters, blogs, photos and more for whatever they're into. And because this is the social web, these feeds can be published to Bluesky or published as social websites that live outside of Surf. Already some amazing creators have launched sites like verge.surf.social, wired.surf.social, filmfeed.surf.social and allnet.surf.social. Social websites represent a whole new way for creators to build destinations for their communities around their content, without being owned and operated by the social media giants of today.

I especially want to thank @chrismessina for not only hunting for us but for being a source of inspiration for how these feeds and sites should work across identities and accounts on the social web. Super excited to hear what you think and answer any questions you have.

btw there's a cool easter egg that people haven't discovered yet. Click on the stats under the surf logo on surf.social.

Happy launch day! The creator-controlled moderation is a really interesting part. How does the tooling for that work? Is it more about blocklists and filters, or something more advanced?

I like how it aggregates a creator's content and allows to filter it in different ways.

As a viewer, am I supposed to be able to interact with the content beyond viewing it? Like how am I supposed to comment. Do I have to go create a mastodon account to comment? And then a bluesky account to comment on the blusky posts? Or does surf handle all of that?

Hi everyone!

Mercury Edit 2 is not a general chat model for coding. It is purpose-built for next-edit prediction, one of the most latency-sensitive parts of dev workflows.

The interesting part is that it is built on a diffusion architecture, so it generates tokens in parallel instead of one by one, which is exactly why it can feel so fast. Inception is claiming 75.6% quality at 221ms, plus a 48% higher accept rate and 27% fewer shown edits than the previous version.

If you use @Zed , there is a specific API key that unlocks a free 1-month trial.

You can find the configuration tutorial here.

Congrats on the launch! A diffusion LLM purpose-built for next-edit prediction is a really interesting angle, and the latency advantage over autoregressive models seems like it could be huge for IDE integrations. Are you seeing the biggest gains in specific languages or is it pretty consistent across the board?

Diffusion architecture for code prediction is a bold bet. Most completion tools just throw a bigger autoregressive model at the problem and call it a day. Curious about the latency in practice though. 221ms on paper vs 221ms when you're mid-flow writing Flutter code are very different things. Does it handle Dart well or is it mostly tuned for the usual Python/JS suspects?

The 'next-edit' framing is interesting - it's predicting intent rather than continuation. How does it handle non-local edits? Like, you rename a function and it needs to chase all the call sites. Is that in scope or is this more single-cursor stuff?

is 'next-edit prediction' meaningfully different from standard autocomplete, or is that just a frame for faster completions?

the diffusion architecture is where this gets interesting. autoregressive models generate one token at a time, so by the time you've generated a complete suggestion for one location, it's too slow to extend across several. diffusion generates token positions in parallel, which makes 'what else changes after this edit' tractable at 221ms. the 48% higher accept rate is the number that actually matters here. low accept rates train developers to dismiss suggestions without reading them. if mercury edit 2 is genuinely better at predicting which edits to surface next, that changes the daily feel more than raw latency numbers do.

Fahhh! Btw have you yourself have become desensitized to the “Faahh” over time?

I've been building a focus timer for 2 years and somehow never considered the "scream at the user" approach. Respect for the pure chaos energy here. Honestly the desk slap as a trigger is lowkey genius because it's a physical action that snaps you back. Most focus tools (mine included) try to be all zen about it. Sometimes you just need to be yelled at.

Hey everyone 👋

I kept opening X “for 5 minutes”…
…and somehow losing 2 hours without noticing.

So I built Faahh — an app that doesn’t remind you…
it interrupts you.

It sits quietly in your menu bar —
then suddenly hits you with a loud “Faahh…” 😄

No dashboards. No blockers.
Just instant awareness the moment you drift.

⚡ What makes it different

• Actually calls you out (not gentle nudges)
• Upload your own MP3s (roast yourself 😄)
• Smart idle detection
• Escalates if you ignore it
• Rewards you when you get back to work

• 🖥️ USB Sentinel — unplug something → your laptop screams
• 👋 Desk Slap Trigger — literally slap your desk to fire a roast
• Runs silently in tray / menu bar
• Works on Mac, Windows & Linux

Inspired by apps like SlapMac — but I wanted more control, customization, and personality.

Built this in 24 hours as a fun experiment…
didn’t expect it to work this well 😅

Curious — which feature would you actually use?
Idle roast, USB alarm, or desk slap? 👀

Really like the idea. I find myself easily ignoring the small nudges, so would love to try something like this. Most interested in the idle "faahh". Can you customize how long before the sound

Great launch Nagaraj, keep shipping 🚀

not gonna lie, this is the kind of aggressive motivation I need sometimes 😂

That's really great! It's excellent for developers.

By the way, I wanted to ask: the tool I used previously—Adspower—seems quite similar to yours. Is it a competitor of yours?

Really cool project! Am I able to run cron jobs to query LLMs with this?

It's super cool Andy! Being able to set different profiles with their own proxies and cookies is super useful for automating daily processes. Wish you all the best here!

Hi everyone!

This is one of the latest experimental projects out of OpenRouter Labs. The concept is kinda similar to the Model Council: you run your prompt through multiple models, a pre-fuse judging step analyzes their outputs across different axes, and then a final "judge" model synthesizes everything into one final answer.

The obvious advantage here is OpenRouter’s massive catalog. You can mix and match almost any current SOTA model available today, whether open or closed.

One interesting variable is the fuse model itself. The model you choose for that final synthesis step seems to have a big impact on the style and quality of the final output, which adds a whole new control layer to the workflow.

Just a quick heads-up on cost: you can definitely select free models to test the waters. But if you route this through premium models, costs can add up quickly, so make sure you have some credits loaded up!

The judge model being configurable is the part that makes this actually interesting. Most "run multiple models" approaches just pick the longest answer or do basic voting. Being able to control the synthesis step adds a real control layer. Have you guys seen meaningful quality jumps when mixing open and closed models together vs sticking with one family?

Most multi-model tools optimize the wrong step. Routing is easy. Synthesis is the hard part. The fact that the judge model is configurable is actually the interesting design decision here, not the parallel execution. A judge that just averages or picks longest is no better than a single good model. Curious whether the default judge behavior is documented somewhere, or if the quality of the fusion is mostly opaque until you've run it enough times to calibrate it yourself.

Running models side-by-side and fusing the best answer is genuinely useful — but it also expands the governance surface. When 3 models contribute to an output, who owns the audit trail? Which model triggered the downstream action?

Microsoft swapped the model inside Copilot this week and most teams had no process for catching it. The teams that get the most from fusion tools like this are the ones who pair it with clear ownership of the decision layer — not just the output quality layer.

As a fellow Mac developer, this makes me really happy. Preview has been frustrating for years and Quick Look disappearing on click is one of those tiny annoyances that adds up. Love the "simple on the surface, obsessively refined underneath" approach. That's exactly the philosophy I try to follow with my own Mac apps. Quick question: does it support HEIC and RAW files natively? Those are the formats I deal with most and Preview handles them poorly.

Here's the homepage:

https://storepal.app/vista

Looks useful — is it possible to replace Preview with this for image files?

Hi folks,

Finally, APImage launched on Product Hunt. We'd love to welcome you on our platform!

As a little thank you, we introduced a launch discount, so that it's easy for you to give it a try.

Your feedback is much appreciated! :)

Best,

Ole

Looks cool! When will the "Teams with centralized billing" be available?

Congrats on the launch! The focus on consistent characters and object preservation across generations is something a lot of AI image tools still struggle with. Are ecommerce product visuals the main use case you're seeing, or is it broader than that?

Hey Product Hunt! founder of Klick 📸

Over the years, I kept getting asked the same thing:

“Can you take this photo for us?”

Friends, strangers, even random travelers — they liked how my shots looked. I run a small side photography page (https://www.instagram.com/xenesfilms/) on Instagram, and composition has always been the thing people notice first.

But here’s what I realized…

Most people don’t struggle because of bad cameras.
They struggle because no one ever taught them how to frame.

I’ve seen it so many times — someone takes a photo, it looks fine in the moment, but later something feels off. Crooked lines. Awkward spacing. Unbalanced framing. And once that moment is gone, it’s gone.

At the same time, AI photo apps were exploding.
But almost all of them focused on fixing photos after you take them.

I kept thinking:
What if the camera could guide you before you press capture?

That’s why I built Klick — an AI camera that gives you live composition feedback while you’re framing your shot. It helps you learn rule of thirds, symmetry, balance and visual alignment in real time — without overwhelming jargon.

What’s inside Klick?

- Live Composition Coaching – Real-time framing guidance while you shoot.
- Smart Framing Modes – Rule of Thirds, Center Framing, Symmetry and more.
- Minimal, Distraction-Free UI – So the focus stays on your eye, not filters.
- Instant Feedback – Know your shot works before you take it.

The goal isn’t just better photos.
It’s building confidence and visual instinct.

We’d love your feedback on:

1. What’s the hardest part about getting a great photo?
2. Which composition styles should we add next?
3. Where does real-time guidance feel most useful — portraits, travel, street?

I’ll be here all day answering questions and taking notes.

If Klick helps you capture even one photo more intentionally, your support or comment would mean a lot.

Thanks for checking it out 🙌

Love this. Teaching composition in real time before you even press capture is such a smarter approach than fixing photos after the fact. Curious how the AI adapts to different scenarios like low light street photography vs bright outdoor portraits?

Looks great! Are you going to make an app for Apple Watch? It could very handy during the photo taking

Please also include this feature for some online movies. 😅

the reason all those one-off utilities ended up as websites is that nobody wanted to hunt down a separate npm package for each use case and pin versions. opengyver's bet is that bundling them under a single cli fixes the discovery problem, so conversion, encoding, uuid gen, json manipulation all stay in the terminal where they belong in shell scripts anyway. the web tab was never the right runtime, it was just the path of least resistance.

What do you think seperates boilerplate and unique, well-made ui? There are obviously so many small points, but what is one big principle?

Just made this on clover in one shot, really great UI

unable to signup

the github export is the right call here. most ai frontend tools make you copy-paste code into your own repo, which means the generated structure is frozen the moment you start editing. giving you an actual repo to clone means the output is designed to be the start of something, not a finished snapshot, and that's a different quality bar entirely.

the tricky part will be the second prompt. the first generation usually looks clean, but when you go back to add a new section or swap out a component, the model doesn't always know what decisions it made before. state management conventions, naming patterns, the way tailwind classes are organized, all of that needs to stay coherent across iterations, not just on the first pass.

How are you thinking about the ownership layer here? With Cursor and Bolt you get code you own - some tools lock you into their renderer. Curious whether this outputs clean editable component code or keeps you in a drag-and-drop environment.

Congrats on the launch Alby and Abdul :) curious, do you guys showcase your favorite frontends generated by Clovr anywhere?

Hey Product Hunt 👋

I'm Shreyak, co-founder of WTMF AI (What's The Matter, Friend?).
Here's why we built this.
1 in 3 Gen Z users say they have no one to talk to. Not "no one close." No one at all. People are journaling into the void, venting to strangers online, using ChatGPT as a therapist because it was the only thing that didn't judge them.
We didn't want to build another utility.
We wanted to build the friend that shows up at 3 AM.
WTMF isn't therapy. It's the AI bestie that remembers your last bad week, picks up where you left off, and doesn't make you feel weird for needing to talk.
Today we're bringing it to Product Hunt. Would love your honest feedback, brutal is fine. Ask me anything.

- Shreyak, co-founder @wtmf

I can see you already have some traffic on the website. How do you make your apps visible to users? (In the meaning of marketing) :)

Super interesting launch fs! What specific signals are you extracting at the AST level that correlate strongly with seniority or depth? Also don't you guys think this system unintentionally favors those engineers who had more free time or OSS exposure etc. which is not bad but still kinda subjective

Hey Product Hunt, I am Tec Narayan Brahmachari, co-founder of GradPipe.
Most AI career tools just polish resumes and optimize for ATS keywords. But startups don’t hire formatting, they hire technical depth. In interviews all they care is if you have the tech-depth or not.

Today we're launching Delta Analysis. It doesn't sugarcoat; it tells you exactly why you aren't ready for your target role and how to fix it

You need to know exactly what the top 1% are actually building.

So I built the Delta Engine into GradPipe.

Instead of generating fluff, it:
• Analyzes your exact goal + your knowledge graph.
• Queries our database of 1,500+ real top-tier engineer resumes to find the exact peers who are beating you and aligned with your goal.
• Shows you the evidence-based gaps (e.g., "Peer 3 built a custom eBPF scheduler, you only have basic CRUD apps").
• Ensures every recommendation is backed by real peer data (the kind of mentor you'd love to have).

It’s powered by our proprietary skill graph.

P.S. If you run into any bugs or issues, or want to share a feedback, shoot me an email directly at tecnarayan@gradpipe.com

Hi, I am Pratham. 4th year at IITB, and one of the Co-Founders of GradPipe.

Delta engine as the name suggest, tells you what's the 'delta' between you and the role you are chasing.

We have the data of 1500+ strong tech builders and AI researchers, already working in companies/labs where you want to be.

It will give you the 'gaps' that exists and the 'roadmap' to fill those gaps. based on the data of the top 5 people who are already pro in that.

Feel free to drop your feedback here. or on the website itself (there is a Share Feedback button on the top right of the delta page)

Hey Hunters 👋

We built NeoWeb because many website builders focus heavily on freedom, but not enough on the quality of the final website.

For a lot of founders and small businesses, the challenge is not only creating a page. It is creating a website that is clear, fast, accessible, search-friendly, and built in a way that follows modern web standards. In many tools, those parts are either manual, easy to miss, or easy to break.

That is the problem we wanted to approach differently.

NeoWeb uses a structured system with fixed components, so users do not have to manage layout complexity like grids, spacing, or design systems. They focus on content, while the platform keeps the output cleaner and more consistent.

A big part of what makes NeoWeb different is what happens behind the scenes:
🔎 SEO foundations are built in
⚡ performance is treated as a core part of the product
♿ accessibility and best practices are kept in mind by default
🧩 structured data and standards-based output help websites be more visible and better understood by search engines and other systems

So NeoWeb is not only about making websites easier to build. It is about helping people create websites with stronger technical quality and better visibility from the start.

We would really love your feedback and thoughts on this approach 🤍

Is customer support available if I get stuck?

How fast can I actually launch a website?

Is coding required at all to use this platform?

Does it support mobile-friendly websites?

Congratulations on your first sale 🙌. You're building solution to a problem I feel regularly on LinkedIn.

Hey everyone! I'm Kevin, the maker of Daydreamer.

Honestly, I built this because I wanted it for myself. I spend a lot of time in front of my Mac, and most break reminders feel like interruptions. I wanted something that actually made me want to take a break. So I made Daydreamer.

I use it every single day. The scenes, the music, the little quotes that show up during each break. It genuinely makes my workday feel a little warmer.

This is my first time launching it publicly, and I'd love to hear what you think. What do you like? What feels off? What features would make it more useful for you? I read every comment and take feedback seriously. There's a lot more I want to build, and your input will shape where this goes next.

Thanks so much for checking it out. Hope Daydreamer brings a little calm to your day!