PH热榜 | 2026-08-19

Hey everyone!

My name is Vida, and I am the Founder and CEO of Astute (www.joinastute.com). 👋

A bit of context on us: I ran growth at Fluidstack through the $2M to $2B stage, @abhishek_manikandan was an ML engineer at The Trade Desk. We met at EF London and just closed an oversubscribed $1.2M pre-seed from VCs and angels to build the first B2B new media company, starting with creator partnerships.

As a B2B marketer, I knew that the marketing playbook (ads, PR, SEO) doesn’t work anymore. At Fluidstack I quickly realised that our audiences weren’t clicking on our ads. Instead, they were reading Substacks like SemiAnalysis, listening to podcasts like Acquired and TBPN, and scrolling on LinkedIn and Twitter. 

Every team I speak to already knows this. But most can't activate this channel today, because it takes months of manual searching, cold outreach, and back-and-forth, with most of it going nowhere. 

Even the current tools in the space don’t work in the way that brings value. As a marketer, and now a founder, I want to understand the best types of creators, the optimal audience size, best performing platforms - for my company. I want a personalized data-backed strategy. And then I just want it all done for me. I don’t mind investing $10k in creators who will bring me customers, but I don’t have the bandwidth to manage the CRM, comms, payments. All the current players fail to understand this.

On top of that, most offer only ad slots, but the real value is in an endorsement or editorial partnership with a creator. But editorial partnerships are difficult to productize. They lead to organic content which is why they require a higher volume of granular data from both sides, to ensure the right match. Competitors haven’t cracked this - we have, with over 16,000 creators in our proprietary network.

So we built Astute. We have two AI agents - one is a new media manager for B2B companies, the other is a talent manager for creators.

How it works

1. Astute first listens. Astute monitors over a million creator posts every minute and surfaces where you, your competitors, and your category are being discussed, including the specific threads worth engaging with.

2. Then it builds a strategy. 1,000s of data points turn into a creator strategy specific to you: who your buyers already follow, which audiences overlap with your ICP, and what a realistic first three months looks like.

3. It handles the partnerships. The manager agent searches the 16,000 creators on the talent side - people with loyal, high-intent audiences on Substack, podcasts, LinkedIn and X - picks the right ones, and locks in the partnerships once you approve. Editorial features, podcast guest slots, newsletter ads, event collabs, and more.

4. And it measures ROI. Impressions, conversions, brand mentions, and visibility inside AI search engines.

But this is only the beginning. We are going after the full new media stack. We want to build a space where every B2B company will come to build their brand, legitimacy, and an owned audience in new media.

Now the juicy stuff - our Product Hunt deal

• Go to https://company-app.joinastute.com/ 

• Apply code PRODUCTHUNT

• Do it before 31 August, 23:59 PT - the code expires automatically

Finally - HUGE thanks to @vedranrasic for hunting us today 🚀

The measurement side is where I'd expect this to live or die, and it's harder than it looks. I built proper attribution on my own site last week and the thing that surprised me is how much creator and social traffic arrives carrying nothing at all: messaging apps, iOS Mail and in-app browsers strip the referrer, so a meaningful share of my signups log as "direct" and I genuinely cannot tell which post caused them. For a B2B buyer the first question is going to be which partnership produced pipeline. How do you attribute that when a lot of the clicks arrive anonymous by construction? If you've solved that cleanly it's a bigger selling point than the automation.

Hey - I'm Abhi, the Founder and CTO of Astute!

Before Astute, after graduating from Oxford Uni (CS), I worked as an ML engineer at The Trade Desk, but on the side I ran a B2B engineering YouTube channel that ended up doing 15M+ views (itisAbhi on YT). I've personally received 100s of partnership emails from brands.

And to be honest, I said no to almost all of them. Not because I didn't want to work with companies, but because saying yes meant weeks of back-and-forth, contracts, chasing invoices... And with a lot of companies I also didn't want to work, because they were not relevant to my content at all. It made zero sense.

So when we started building Astute I was really excited to fix that side of the problem. The creator side. We built the talent manager agent I wish I had. And over 16,000 creators in our network are already using it!

Happy to go deep on how any of it works - the matching, agent architecture, whatever. All questions are welcome!

Hey all - Founding Engineer @ Astute here!

When @abhishek_manikandan and @vida_stanic1 told me what they're building - I knew I had to get on this. My family runs a big media company in my home country - so I REALLY get this.

I've been working mainly on the social listening and media monitoring element of the platform, which all powers our data strategy. Feel free to shoot over any questions you may have for me :)

Super excited to hunt this one.

I run LeadDelta (PH SaaS of the Year 2022), so I live inside B2B distribution all day. Somewhere in the last 2 years, the answer to "where did this deal come from?" stopped being an ad. It's a newsletter now. Or a podcast. Or someone on LinkedIn my buyer has been following for three years. Everyone I talk to knows this is happening.

So when Vida showed me Astute I got it immediately. 

The more I got to know her and her co-founder, the more I realized there isn’t a better team to build this company. Vida spent years at Fluidstack helping grow the company from $2M to $2B using this exact channel! Her co-founder Abhi was a creator himself - 15M+ views on YouTube is HUGE - he's literally the guy on the other end of the partnership emails.

Vida and Abhi are around all day, ask them anything. 

Congrats on the launch team! Proud to be a customer (and Hunter!).

It was a WILD RIDE working on Astute with @vida_stanic1 and @abhishek_manikandan, especially the past few weeks leading up to our launch!

I am so excited to share what we've built with the world - it's really a game changer for Founders and Marketers!

Feel free to @ me with any technical questions :)

Super cool! Love the direction. In the past few years, I noticed that all kinds of marketing simply repel me, and I don't trust the product. A lot of the ads and SEO are AI slop anyway. But when I see someone I read, e.g., Lenny or Pragmatic Engineer, share some tool, I instantly get a thought: "Wow, why haven't I tried this before?"

Congrats on the launch!

@vida_stanic1 Congrats on the PH launch and the $1.2M raise.. Anyone who has tried to run creator partnerships manually knows how much of it is chasing contracts and invoices.

The amount of time as founders we’ve spent trying to go viral is unreal. Perfect team for a massive problem, congrats guys!

As a creator, this immediately caught my attention. Excited to see brands and creators connect in a more meaningful way!

Congrats on the launch team! Curious to understand as Astute expands beyond creator partnerships, what other parts of “new media” do you see the platform helping companies manage?

Big congrats on the launch! Automating B2B brand virality through new media creators is such a powerful angle given how fast distribution tactics are shifting right now. Love seeing tools that help founders tackle creator-led marketing without the usual operational nightmare. Wishing the team massive momentum today!

Looks super exciting!! Congrats on the launch, will send to my team

See you there for the 2nd time. You are nailing your distribution! :D :)

Congratulations on the launch of Astute! (and a terrific launch video!) Compared to B2C, B2B has a lot of catching up to do due to the shift in marketing from the old playbook to how audiences now consume media. Very excited to watch your journey! @vida_stanic1 @abhishek_manikandan

Congrats on the launch, @vida_stanic1 and @abhishek_manikandan !

The B2C market has already proven the creator economy is massive - I actually believe the B2B opportunity is even bigger, there just hasn't been a tool that makes it possible until now.

That's exactly why Astute clicked the moment I saw it. Excited to watch this take off.

This is a problem I’ve run into countless times myself: you know the audiences you want to reach are listening to podcasts, reading newsletters and following creators, but actually finding the right people, figuring out who to work with and managing everything is incredibly manual.

So excited about what @vida_stanic1, @abhishek_manikandan and the @Astute team are building. A problem I genuinely wanted solved. Huge congrats on the launch 🚀

As a creator, this immediately caught my attention. Excited to see brands and creators connect in a more meaningful way!

Congrats on launch @Astute 🚀
Interesting! Does this work well for startups too? Do you have any special packages for startups?

Pretty cool! Congrats on the launch@vida_stanic1 and the Astute team. Love the video!

Yep, putting right companies in front of a right audience with the relevant message - might sounds simple but the devil is in the details, as usual. And such a massive pain - already registered!

Congrats! Really excited for this

Really interesting approach! Turning creator partnerships into a data-driven, automated process for B2B sounds very promising. Love the idea of making something that usually takes a lot of manual work much faster and more scalable.
Good luck guys!

Experiencing this problem right now, trying to find + manage specific creators for a campaign. Can't wait to try!

Hi Product Hunt! 👋 Henry here, a few months ago, I launched @Clipto , a fully local AI media search engine that helps people search terabytes of videos, photos, audio, meetings, and documents using natural language.

Since then, one request kept coming up:

“Can Claude use Clipto?”
“Can ChatGPT search my media library?”
“Can my agent actually edit videos using my own footage?”

Today, we’re excited to launch Clipto MCP.

AI agents can already access your files. What they can’t do is understand what’s inside terabytes of media.

That’s what Clipto MCP changes. It gives AI agents semantic understanding of your local media, so instead of manually browsing folders or scrubbing through timelines, you can simply describe what you want.

Some examples:

🎬 Turn a script into a video by matching every sentence with relevant footage.

🔍 Find every clip where someone mentioned a specific topic.

🎤 Search years of meetings for a decision or discussion.

✂️ Generate rough cuts from thousands of hours of video.

Everything runs on the media already stored on your computer. No uploading your library to the cloud.

This is only the beginning. As AI agents become more capable, they’ll need more than file access. They’ll need to understand the content inside our personal media.

To celebrate our launch, we're offering 1 month free to anyone who signs up this week with code PHLNCH.

We’d love to hear what you’d build with Clipto MCP.

We’ll be here all day answering questions and collecting feedback.

Thanks for giving it a try! 🚀

Really curious how well the semantic search performs once you throw a huge media library at it.

congrats for the launch! this is really helpful, I can imagine starting a project in Claude and asking it to find all the relevant material before I even open my editor 👀

congratulations! what has been the most surprising use case you have discovered while testing agents with large personal media libraries?

Meeting recordings might be the boring use case here, but honestly probably one of the most useful.

I’d be interested to see this with a really messy archive — random filenames, old projects, mixed audio and video.

Does this work the same way with Claude and Cursor?

I can see this fitting really naturally into an editing workflow. Find the right clips first, then build from there.

Looks cool! How well it handle the large media libraries ? Does performance stay consistent when searching through tb of files ?

The “find this moment in my footage” use case is probably the first thing I’d try.

This could be useful for research too, especially when your source material is mostly video and audio rather than documents.

Congrats @henrykang and the team on the official PH launch of Clipto MCP! 🎉 I’ve been following since your first release, and this MCP upgrade is a game-changer – the idea of letting AI agents "understand" terabytes of local media instead of just accessing files is brilliant. "Turn a script into footage" and "search meeting decisions" are killer use cases for creators and teams alike.

Huge props for keeping everything 100% local (privacy first!) and that 2TB/24hrs indexing speed on M5 is seriously impressive. 💪

One concrete suggestion: since MCP is all about agentic workflows, could you add semantic filters like "emotional tone" (e.g., excited, serious) or "shot type" (close-up, wide) for more nuanced retrieval? Also, custom tag hierarchies (project + client) would make enterprise adoption much stickier.

Quick question: any plans to support shared indexing across NAS or external drives? Or maybe a lightweight API for developers to plug into tools like Notion? I'd love to see how far this ecosystem can go.

Wishing you a huge launch! 🔥 Can't wait to hear your thoughts!

The Elon/Daft Punk demo is wild lol.

How much manual cleanup did you actually skip vs what Claude + FFmpeg pulled off on its own?

Documentary editors sitting on hundreds of hours of footage are probably going to appreciate this one.

The local search part is what caught my attention. Being able to ask an AI tool to find something across your own media library feels genuinely useful.

Such a smart idea that addresses a real pain point. The hardest part of putting together a demo video is looking for that particular screenshot called "Screenshot 2026-08-19 at 12.04.44 PM" or video named "858551D6-F8A4-4CE0-9A71-DC6EF6DDD2A5" 🤦🏻‍♀️

I'm excited to try Clipto's NPL search for this and curious if the tags work!

This feels especially useful for projects where the context is spread across dozens or hundreds of files.

Congrats on this launch! Amazing product!

I like the new use case for the application! Because photos are easier to store than ever, I find it harder to find what I'm looking for in my camera roll than before.

Congrats on the second launch. That music video went out as the raw first output, no cleanup at all. Takes some nerve to show the actual thing rather than the polished version.

This is so cool, I desperately needed something of this sort for so long

Sourcing clips from local video terabytes using AI agents is a huge time-saver. Congrats on bringing this to life!

I’d love to see more examples of Claude or Cursor taking the retrieved content and turning it into a finished project.

As someone with TONS of footage, not the best file organization (to be kind to myself about it) and content creation as a side project, I am really excited to try this one, congrats on the launch!

Pretty cool and handy. Keeping everything local is a big plus.

The demo where the agent finds the relevant material and keeps working from it was probably the most interesting part.

Being able to search across local video and audio through natural language could save a ridiculous amount of time.

The editing example got me thinking about how much time is wasted just hunting for the right shot before the actual work even starts.

Being able to search across local video and audio throughnatural language could save a ridiculous amount of time

Excited to see this launch, team! Definitely arrives at a necessary time.

We're working on an integration to allow automatic deployments of sites and functions on Appwrite from Origin, would love to connect with any relevant engineering/DevRel team members about the same.

Can't wait to test it!

Really cool. Congrats on the launch, Justin. Most of what my team ships internally is agent-written now and the review step is the part that hasn't caught up, so that's where I'll be testing this.

@GitHub down earlier this week, the timing is perfect.

Love the new beta stacked feature in github, as long as you dont need to rebase :D. How does this stacked feature in the new cursor origin compare?

I'm now a daily user of github stacked PRs feature, excited to try this.

Eleven launches in and you are still taking the hard route. Congrats on building a whole Git forge instead of another editor feature. That is a lot of work for a company that could coast.

My entire codebase is now officially cursor-native: cursor + grok-bot + origin and bugbot

Love that Origin puts the repo right next to the agents working on it instead of treating agent workflows as an afterthought bolted onto a human-first forge.

Hey Product Hunt 👋 I built this after the Claude watermark news in August, when the internet filled up with tools promising to remove a watermark nobody outside Anthropic can actually detect. This does the part that is real. Copying out of a chat interface carries things with it: HTML class names, zero-width characters, exotic spaces, typography. Paste anything and you get every one of them with a count and a position, because those are facts about the bytes rather than a probability score. One click strips them. Two things I would rather say up front than bury. First, it does NOT detect Anthropic's statistical watermark. Nobody outside Anthropic can, because verification needs a key that has not been released. Any tool claiming otherwise is guessing. Second, the em dash is not the tell people think it is. I ran ten pre-computer novels through the checker: Melville uses 26 per thousand words in Moby Dick, while Austen and Stoker use none at all. A signal that swings from 0 to 26 across human authors cannot accuse anyone of anything. The detection engine is MIT open source, so you can check both of those claims instead of trusting me. Free and unlimited to check, runs in your browser, nothing uploaded. I would genuinely like to hear where it gets things wrong.

Cool approach, and agreed that you can't detect Anthropic's watermark without their keys. But on removal I wouldn't fully agree: you don't need to detect a watermark to remove it. There's research showing these watermarks mostly don't survive paraphrasing or translating back and forth.

You can't use "Claude" in the name, ban incoming.

This is such a great response to their updates :D

You ran ten pre-computer novels through the checker to test the em dash theory. Moby Dick hits twenty-six per thousand words, Austen none. Congrats on shipping something that leads with what it can't do.

I don't see any issue with cloud watermarking, but anyway, good niche findings! Good luck!

As a former marketer, I loved the benefits of AI agents, but hated the complex setup and the fact that my second brain needed extra steps to be plugged in.

So we decided to change that with Cluing.

Cluing agents are fully hosted with us, require no technical setup, and learn right alongside you. Custom connectors make sure whatever your tech stack is, we've got you covered.

You can even take us with you and manage agents straight from your phone.

Can't wait to hear your thoughts 💛

Hey Product Hunters, miss us? 😄

We just launched Autonomous Agents in Cluing - Claude Code capabilities, hosted and shared across your team! I've spent the last few weeks basically living inside these agents, so I'm genuinely thrilled they're finally out. 🎊

A few things I love about them:

🧠They already know your context - your knowledge and past work live in the same workspace.

🔌 If a connector doesn't exist, the agent just builds one (and stores your credentials securely).

⚒️ Skills stay in sync with your Topics, so your whole team stays aligned without rewriting prompts.

🌐 They can publish a page or app live, right from Cluing - even on your own domain.

👥 And they live in a shared workspace, so your team works with the agent, not around it.

Best part for a non-developer like me: if you can describe a task, you can run an agent for it 😄

I'll be around all day and would genuinely love to hear what you'd connect an agent to first. Ask me anything! 💚

Long time no see makers! Tech Founder here 👋

The idea for hosted agents came from our team, we were using different harnesses and automating our work but:

➜ The mac needed to stay open - Cats would also shut down our agents

➜ Real collaboration was almost impossible

➜ I spent a ton of time just keeping everything up and running

➜ No true isolation - All agents same machine, running 15 chrome instances eating up the RAM, disrupting one another

➜ And the thing that drove me nuts - Sessions handling was on a harness level - Every new comment would reread things again and again, hallucinate mid-work..

So we started with our own hosted agents inside our team, the Cluing memory was natively already there, after we got a ton of value out of it, we decided to make it a part of the product for everyone.

There's a 25% OFF all seats, there's $25 Free Credits ($300 for yearly subs) for every new subscription, give it a try!

Very exciting, congrats on the launch!

@ralic @gordana_laskovic @nemanja_trtica congratulations on the launch!

Really like the direction here. Moving agents from something that lives on one person’s laptop to something the whole team can actually collaborate around feels like a big unlock.

Curious, what’s the most impressive end-to-end workflow you’ve already seen a team hand over to a Hosted Agent? Would love to hear what it can do in the wild.

congrats team 🎉 the no-setup part is what makes this click for me. rooting for you

Needed a solution for my current problem. I’ll be trying it out. Congrats on the launch!

Cats kept shutting down your agents on the old setup. That's the most honest reason to host anything I've read today. Congrats on the eighth launch, four years in and still writing like people.

Hey Product Hunt! 👋

I am J.R., creator of Paper Critters.

Paper Critters lets kids design custom paper toy characters using full sticker controls, then print and fold them into real physical toys. It bridges screen time with hands-on play, making it a great creative outlet for kids, parents, and educators.

Key features to check out:

• Kid-Safe Privacy: No email collected, no account required to try, and built around child privacy constraints from day one

• Frictionless Web App: Fully responsive Progressive Web App that installs to your home screen with no app store download required

• Safety First: Human moderation on all public gallery submissions

• Monetization: Free to design and explore, with a simple one-time purchase to print high resolution templates with cut and fold lines

I would love your feedback, especially around the sticker decoration controls and the overall print-to-assembly experience.

Thanks for checking it out! 🎨

Super cool toy for kids! Love this

Thanks for using Expert Chase.
Version 2.0, first wave of the redesigned UI

• Reminders from chat on device. Say “remind me to call Mom at 3,” “remind me 30 minutes before my meeting,” “remind me every morning to take vitamins,” or “remind me to drink water.” Works for tasks, events, habits, or a wellness check-in
• Refined Explore and glass-style navigation in the app
• Faster, more reliable everyday interactions
• Quicker inline compose when you add something new
• Stability and performance improvements
• Squashed some bugs along the way

We’re glad you’re with us.

I’ve been building Expert Chase for a while, and 2.0 feels like the version I originally had in my head.

This isn’t just a feature update. I redesigned the experience from the ground up and brought together the parts of everyday life you need in one place.

I’d love for you to try it and tell me what you think.

Expert Chase 2.0 is now live on iOS, Android, and Web.

Download on the App Store:

https://apps.apple.com/us/app/expert-chase-ai-for-life/id6766313000

Get it on Google Play:
https://play.google.com/store/apps/details?id=com.expertchase.app

really interesting idea! curious how you hope to capture users who are currently operating in a more - analog capacity to this? Or is that even part of the prospective demo?

Happy to see you re-launching :)

@eye Congratulations. And happy product launch.

Hi everyone!

One of my favorite @Fairphone details is still the screwdriver in the box. You’re actually expected to open the phone!

The 6+ keeps the same repairable chassis with 12 user-replaceable parts, a removable battery, a five-year warranty, and software support through 2033. The update is mostly inside: Snapdragon 7s Gen 4 and 12GB of RAM, giving a phone meant to stay around for years a bit more headroom.

The “Plus” is deliberately modest. No bigger display or extra camera. Fairphone has basically taken last year’s 6 and made it more comfortable to age.

And for the first time, Fairphone is selling directly in the US.

A phone that ships with the tool to keep it alive for years is a pretty refreshing idea.

Thanks for sharing. It's so rare to see privacy related products here.

Love this idea, my iPhone battery is always dead and here I can just swap it.

We don't have a support team and all - we don't know if this post will be ignored among other launches.

But, for everyone that does land on this page - We will treat EVERY. SINGLE. ONE of your comments with FULL hearts.

Please let us know what you think, it'll be a pleasure.

deploying agents are very difficult for non-technical people, really appreciate omnivibe to make using agents more friendly to us

The Slack-like messaging approach for chatting with multiple agents is my favorite. Congrats team! 👏

There are so many cool agents buried in GitHub repos that most people will never find. I can see the appeal here. Congrats!

Really like the idea of giving AI agents a place beyond scattered GitHub repos and local setups. Making them easier to discover, use, and even monetize feels like a natural next step for the agent ecosystem. Congrats on the launch!

I appreciate that the product does not assume every user knows what a SKILL.md file is. The best agent experience should feel like using a product, not configuring a developer tool.

This is huge, congratulations on the launch! I know @lotus_lu1 personally and she is a GTM guru. If some other teams building an agent marketplace product, I will be suspicious about the distribution. With Lotus leading the GTM, I am sure OmniVibe has a high chance to success.

Keep going!

The biggest blocker for me is discoverability and trust, knowing people will actually find it and that it's safe to use. A marketplace with good curation would solve both.

Hey PH,

I often want to automate recurring tasks with my Claude Code/Codex subscriptions, so I built Cronloop to make this simple.

Just click "New agent", set a schedule, connect your existing claude code/codex subscription, give your agent instructions, connect the tools it needs, and boom - you have an autonomous agent running in a loop, working while you sleep.

Some ways I'm using it:

- I built a self-driving events website (aievents.now) and have a cronloop agent for each city that autonomously curates events every morning.
- I have a few websites autonomously SEO-optimizing every day based with connected search console data + keyword research tools - the site runs on Cactal.ai.

Some features:

- Every agent runs in an ephemeral, isolated sandbox
- You can monitor runs in real-time (the agent logs stream into the app)
- Every agent has a simple, durable markdown-based memory system. Your agents can self-improve over time - you can include instructions to record learnings with each run for other agents to benefit from.
- You can leverage your existing claude code/codex subscriptions or bring your own API key.
- 200+ connectors are ready-to-go - connect your agents to any tool that has an MCP, CLI, SDK, API, etc.
- You can configure the agent's environment with a setup shell script if needed for greater flexibility

Free to try - create up to three agents for free.

You bring your own inference, so cronloop is very generous with usage, the pro plan offers unlimited agents with up to a 5 min interval on runs for $25/mo (or $20/mo annual).

I'm a huge Cronloop user and get a lot of value from it (I personally have ~25 cronloop agents running), so I'm excited to share it. Let me know your thoughts!

*Hand-written comment, no AI

Cute fun idea! It would be really cool if it would be an easy to install extension for different browsers (for non-techy people like myself).

Hey Product Hunt, Praney here. I build Vois solo from Melbourne.

Since the first Vois launch here, one complaint kept coming back from
podcasters and narrators, and I had it myself: cloud voice tools charge per
character, so every retake of the same paragraph costs money again. I
started calling it the preview tax. You stop iterating, and the read stays
mediocre.

Vois 2.x is my answer: an ElevenLabs alternative that runs on your own
machine. Every subscription has unlimited generation: no tokens, no usage
meter. Retake a paragraph fifty times and your bill does not move.

What is in 2.x:

- Four engines, picked per job: Fast for drafts (about 3x real time on CPU,
about 6x with GPU acceleration on Apple Silicon), Expressive for
delivery, Multilingual for 23 languages, and Omni on Pro with 600+
languages and Voice Design.
- 100+ voices across 21 categories, so a recurring narrator stays
consistent.
- Voice cloning from about 15 seconds of clean audio, with a consent step.
- A multi-speaker timeline with music, mastering, and one-click export.
- Vois CLI for AI agents: Claude, Codex, Cursor, or any agent that can run
a shell command can create projects, generate batches, and export while
you review the plan. Your agent does the talking.
- Since 2.0: encrypted portable backups, reconstructible generation
records, and accessibility across the whole studio.

Scripts, voice samples, and generated audio stay on your machine.

A peek at what I am building next: Quality Guard, which transcribes every
take, compares it against your script, flags misreads, and can regenerate a
bad segment and keep it only if a recheck passes. It is behind a feature
flag today.

For launch week: Subscriber is $10/month and Pro is $14/month(first 100 spots), applied
automatically at vois.so, and you keep this price for the life of your
subscription. Ends August 31, limited spots.

People use Vois for audiobooks, podcasts, faceless YouTube channels, game
NPC voices, tutorials, and course narration.

What would make this fit your workflow? Podcasters, audiobook narrators,
game devs, and anyone wiring voice into an agent pipeline, I am especially
keen to hear from you. I will be here all day.

Best of luck🙌 on Product Hunt today, @praney_behl the vision for privacy-first unmetered desktop AI tools is the exact direction the market needs, can users fine-tune pitch, emotion, or vocal stability on a per-sentence level within the Expressive engine?

Congratulations on the release. I just picked a promo code. Thanks, great work. The Ui is simple and I like the ability to connect to Claude for AI automation.

Hey Product Hunt 👋

I’m building KiHub because there’s still a big gap in the way hardware changes get reviewed.

KiCad projects can live in Git and move through GitHub, but reviewing what actually changed is often surprisingly manual: opening KiCad locally, comparing revisions yourself, passing screenshots around, and discussing design decisions across GitHub comments or Slack.

KiHub adds the missing review layer around that workflow.

With KiHub you can:

• Compare schematic and PCB revisions visually — see design changes in the browser instead of interpreting raw KiCad file diffs.

• Start reviews automatically or on demand — from a GitHub pull request or by comparing any two branches, tags, or commits.

• Discuss changes in context — leave threaded comments pinned directly to a schematic sheet or PCB location.

• Run hardware-aware checks — including ERC, DRC, schematic-to-PCB parity, BOM completeness, and release readiness.

• Set review and merge gates — decide which checks are informational and which must pass, and use review status and hardware checks to gate GitHub merges.

• Review BOM impact — compare component changes, find missing data, and track order readiness.

• Share reviews outside the workspace — publish read-only links for teammates or external reviewers.

• Keep every decision traceable — preserve compared revisions, check results, discussions, decisions, and activity in one place.

KiCad stays the editor, and GitHub stays the source of truth. KiHub is the shared workspace where the hardware review happens.

If you’d like to explore it without connecting a repository, we have a gallery of live, read-only reviews generated from public KiCad projects:

👉 Explore the KiHub demo gallery

I’d especially love feedback from people working with KiCad or reviewing PCB designs today: how does your team currently review hardware changes, and what part of that process is the most painful?

Thanks for checking it out!

Love seeing better tooling for hardware collaboration. Keeping PCB reviews, discussions, and approvals in one place without changing existing Git workflows feels like a practical improvement.

this is basically the shadcn model (source code delivered into your repo instead of an npm black box) but pointed at agents instead of humans copy-pasting, which feels like the right instinct given how much agent-written UI code I've seen drift from whatever design system it started from. the question I'd actually want answered before adopting it: once an agent has extended a delivered component and the registry ships an update upstream, what's the story for reconciling the two? shadcn's answer is basically "you're on your own, diff it yourself" - curious if Balsa has anything smarter than that or if it's the same tradeoff.

The AI cheating problem in hiring is only going to grow, so building trust back into interviews feels like the right direction. Love the focus on evidence over polished resumes.

The real win here is giving non-technical teams more independence without making data teams answer every small question. Great use case for AI.

Love it Kim and even more when I read how it turn into a real product. Wish you all the best here!

This solves a very real creator problem. Turning static case studies into something more engaging without opening After Effects is a big win.