PH热榜 | 2026-08-18

Hey Product Hunt 👋 We’re excited to introduce Clara - an AI SDR by TruGen AI.

For the past few months, our team has been heads-down building Clara with one goal in mind: helping companies turn website traffic into qualified pipeline.

We kept seeing the same challenge - companies invest heavily in bringing visitors to their websites, but most websites still rely on forms, chatbots, or waiting for a sales rep to follow up. High-intent visitors don’t always wait.

That’s where Clara comes in.

🤖 What can Clara do?

✅ Engage website visitors in real time
✅ Qualify leads based on their intent
✅ Give personalized product demos
✅ Answer questions and handle objections
✅ Book meetings automatically
✅ Integrate with your CRM and existing sales stack
✅ Work 24/7 across major languages

Clara learns from your product docs, pitch decks, FAQs, and sales content, so every conversation is tailored to your product and your customers.

🚀 The result?

More conversations with the right prospects.
More qualified leads.
More meetings from the traffic you already have.

No forms. No waiting. No extra headcount.

Clara runs on your website 24/7, giving every visitor the opportunity to have a sales conversation whenever they’re ready.

We built Clara as part of our vision at TruGen AI to create AI teammates that can actually work alongside your team - not just answer questions.

👉 Try Clara: https://clarasdr.ai/

We’d love to hear your feedback, thoughts, and ideas as we continue building Clara. 🚀

What‘s interesting here is the shift from “capture the lead and follow up later” to trying to complete more of the sales motion while intent is still live.

Forms are basically designed for handoff. @Clara AI SDR is betting that the website itself can become the first sales conversation.

The open question for me is where buyers draw the line between what they’re happy to do with an AI SDR and what they still want a human for.

And nice to see @Deepgram under the hood!

@bhavyasree congratulations on the launch! Out of curiosity, how is Clara different from other "lead mapping" solutions like Leadoo?

Turning website traffic straight into qualified pipeline is a massive win for founders. Huge congrats on the launch!

Congrats @bhavyasree ! This is really cool. Having an SDR available 24/7 is a huge advantage for sales teams.

How does Clara answer questions? What information does she use and how many languages can she speak?

congrats on the launch @bhavyasree @hemantha_vijay1 . Good to see you folks are trending at #1.

curious, how are you thinking about measuring Clara’s performance beyond meetings booked? would love to know what signals you use to decide whether a conversation is actually a qualified opportunity.

Hello Product Hunt community! 👋

I’m the CTO & Co-founder at TruGen AI, and for the past few months, our team has been building @Clara AI SDR with one mindset: every website visitor is an opportunity and shouldn’t go unattended.

Instead of making visitors fill out forms or wait for a sales rep, Clara engages them in real time, understands their intent, qualifies them, gives personalized demos, and helps move the conversation forward.

We’re finally excited to share what we’ve built with the Product Hunt community! 🚀

👉 Try Clara: https://clarasdr.ai/

Would really appreciate your feedback and thoughts as you check out Clara.

Thanks so much for the support! 🙌

Lovely launch video. Good luck on the launch

Congratulations on the Launch!

Always happy to share what we’ve been building with the Product Hunt community : )

We’ve been working on making website conversations more meaningful - helping teams engage visitors, understand their intent, give personalized demos, and turn those conversations into real sales opportunities.

👉 Try Clara: https://clarasdr.ai/

Looking forward to hearing your feedback and thoughts! 🙌

Congrats on the launch Bhavya 🙌🏾

Reading this from South Africa and cheering the TruGen team from the sidelines. The framing of "the website itself as the first sales conversation" (Zac's line, but you built the thing) is the shift a lot of founders are still catching up to. Wishing you a huge launch day and week🚀

the real-time qualifying part is the hard bit. how does clara tell someone who's genuinely engaged from someone just clicking around to see what happens, is that off timing/behavior or something in how they respond?

Hello Product Hunt community! 👋

I’m the CTO & Co-founder at TruGen AI, and for the past few months, our team has been building @Clara AI SDR with one mindset: every website visitor is an opportunity and shouldn’t go unattended.

Instead of making visitors fill out forms or wait for a sales rep, Clara engages them in real time, understands their intent, qualifies them, gives personalized demos, and helps move the conversation forward.

We’re finally excited to share what we’ve built with the Product Hunt community! 🚀

👉 Try Clara: https://clarasdr.ai/

Would really appreciate your feedback and thoughts as you check out Clara.

Thanks so much for the support! 🙌

Always happy to share what we’ve been building with the Product Hunt community : )

We’ve been working on making website conversations more meaningful - helping teams engage visitors, understand their intent, give personalized demos, and turn those conversations into real sales opportunities.

👉 Try Clara: https://clarasdr.ai/

Looking forward to hearing your feedback and thoughts! 🙌

Hey Product Hunt 👋 Austin here, founder of Taku.

Taku is changing how people discover and use AI, the way TikTok changed how we watch video.

Before Taku: install Codex, hunt down the right tools, write your prompts, set up APIs, iterate, repeat. Then babysit the whole thing.

With Taku: the best setups find you. Copy from the pros, done.

I previously cofounded Sapient Intelligence that raised over $22million in seed funding, and assembled a team from frontier labs like, Anthropic, Deepmind, Deepseek etc.

I was building a coding agent at the time and I realized AI wasn't making my life easier. It was making it harder. A new tool every week. New prompts, new configs, new workflows to babysit.

And me? I'm a lazy potato. I want shortcuts. I wanted to copy, or honestly just steal, from people who know their shit. That one lazy thought resonated with more people than anything I'd actually built. So we built Taku.

For the everyday Joe: copy a pro's setup and it runs out of the box. From there, Taku learns from every interaction and recommends the tools pros built for exactly what you need. The more you use it, the better it gets.

For creators: you've pumped out an incredible wave of skills and open-source projects, but getting noticed is hard, and getting paid is even harder. Taku turns your work into user-friendly AI apps and recommends them to people at the exact moment they need them.

Give it a try and tell us how to make it better.

What happens when one capability inside a Stack breaks or changes? Can I inspect the failing step without rebuilding everything?

I spent some time trying Taku today, and the part that clicked for me wasn’t simply “another way to build AI workflows.”

There are already a lot of powerful agents, skills, open-source projects, and workflows out there. My problem is that I discover them constantly, save them, and then almost never get them into something I actually use.

What feels different here is the abstraction. Instead of asking me to understand the underlying setup first, Taku tries to turn those capabilities into something that behaves more like an app: I can discover something that already works, run it, then change the goal/context and build on top of it.

I also like the Stack direction. Most useful work isn’t one prompt — for me it might be research → analyze → rank opportunities → turn the best one into a brief. Being able to keep that as one reusable workflow is much more interesting than jumping between five different AI tools every time.

I’m curious about one thing as the marketplace grows: how will Taku decide which apps or Stax to recommend when several could solve the same job? If the recommendation layer gets good enough to understand what I’m working on and surface the right capability at the right moment, that could be a very strong part of the product.

Congrats on the launch — excited to see where the app/creator ecosystem goes from here.

"copy from the pros, done" is the sharp bit here. when two pros have opposite setups for the same job, does taku surface both or just converge on whichever's more popular?

This is very helpful, what kind of integrations do you have?

Hey Product Hunt 👋 Emily here, CMO at Taku.

I joined Taku for a pretty simple reason:

I had access to more AI than ever, and somehow I was still barely using any of it.

I’ve spent my career on the business side — marketing, growth, and building companies. I’m constantly having slightly unhinged ideas and thinking, “wait… could AI build this?”

So naturally, I became an AI tool hoarder.

Codex. Claude Code. GitHub repos with thousands of stars. Every “you NEED to try this agent” thread on X.

My bookmarks looked incredible.

My actual workflow? Basically unchanged. 😂

Because somewhere between “this looks insane” and “I’m using it” came repos, dependencies, configs, API keys, and errors I had no business debugging.

Then Austin showed me Taku, and the idea clicked:

What if all this incredible AI capability felt like apps instead of infrastructure?

I don’t think the next billion AI users are going to learn GitHub, manage skills, or debug environments.

They’re going to expect AI to work like every other great piece of software: find what you need, open it, and get something done.

And on the other side, builders are creating an insane amount of intelligence that still struggles to reach the people who need it.

That’s the opportunity I see in Taku:

turn what AI builders create into intuitive apps everyone else can actually use.

Builders get distribution.

Everyone else gets superpowers without the setup.

That’s why I joined.

Give Taku 2.0 Beta a try — and if anything still makes you feel like you need a CS degree, please tell me. 😭

when you copy a pro's setup, is it a snapshot or does it stay linked to their version? curious what happens when the original creator pushes an update after you've already customized your copy

Making AI user-friendly is genuinely helpful, especially for those who are not so tech-savvy but could do with the AI help

Big congrats on the launch! "Borrow the best AI setups and make them yours" is such a clever hook because curation and workflow-sharing are huge right now for builders trying to skip the learning curve. Love seeing tools that democratize productivity and leverage community setups like this. Wishing you tons of momentum today!

I think the creator ecosystem could become a big part of this. If I build one great workflow, I would rather let other people run and remix it than explain the setup 50 times.

An amazing peek about how we imaged AIOS should be

Congrats on the launch. You said the lazy idea of copying someone else's setup got more of a reaction than anything you had actually built. Good on you for listening to that.

I have a ridiculous number of AI tools bookmarked and probably use 5% of them. Turning the useful ones into something I can just open and run feels like the right abstraction.

As a designer (zero coding), this is the best AI desktop app I've used. I am not a coder at all, I've lost count of the hours spent reading instruction.md files on GitHub trying to make "agent skills" work, and Taku turned that all into just a few clicks.

Made an album app in like 10 mins, more design skills and bundles please 🙏

Very interesting idea! I know I get overwhelmed with the stops to try all these new AI projects, and Taku is making it easier!

I'm trying it out now and want to look at specific use cases for this platform. Can't wait!

I would use this for all the workflows I currently have scattered across ChatGPT chats, docs, saved prompts, browser tabs, and random tools I forgot I signed up for.

tried this and loved it.
i do a lot of pm work and most ai tools die at the setup step for me, i'm not going to clone a repo and hunt for api keys on a tuesday. this one just opened and worked, and changing it to fit my own workflow took a minute. simple in a way i wasn't expecting.
super cool cpncept and i hope more people find out this app!

Love the concept of turning complex agent workflows into actual desktop apps that non-tech folks can use without wrestling with GitHub setups. The "remixing setups from pros" feature sounds like a game changer for getting started fast.

Quick feedback: The UI looks super slick, but having a quick 30-second video demo right at the top of the site would make the value proposition click even faster.

Question: Since Taku learns and brings the right tools over time, how does it handle data privacy and local context for sensitive workflows?

Wishing you guys massive success today! 🙌

the creator side is the part I'm most curious about. you said getting paid is even harder than getting noticed, and Taku turns work into apps and recommends them at the right moment - but what does the actual payout look like for the creator when someone runs their setup, is there a rev share per use, a one-time fee, or is discovery itself the whole value prop for now

An agency founder told me last year: "We're never changing our QA process. It's tight."

I asked what tight looked like. He sent a spreadsheet with 400 checks on it. A ten person QA team. Every site clears all 400 before a client sees it. That's ten people doing nothing but checking, all year.

Two weeks later he came back. He had a mandate to find AI tools for QA.

Both things were true at once. Nothing about the process was broken. It just stopped fitting the number of pages.

Creating got fast and nearly free. Checking costs exactly what it always did.

So we rebuilt Superflow around checking.

You upload the checklist you already own. A spreadsheet, a doc, a PDF. It becomes a team of agents, and each knows what to look for: broken links, a wrong phone number, old pricing, a claim your legal team hates, a page that breaks the brand guide.

They run in parallel across the site, phone and desktop, and pin what they find on the page itself, screenshot attached. Ten agents across 80 pages is 800 checks in a few minutes.

Then the human part. You look at what they found and decide. Reject a finding and it stops coming back. Nothing ships until a person says so.

Free to start. 500 credits at signup, enough for one real site run. Installs on Webflow, Framer, WordPress, Shopify, or a script tag.

I'm here all day, and here's what I actually want. Post the check on your list you think no agent could do. I'll run it on your site and report back, pass or fail. And for everyone who doesn't ship websites for a living: what do you always spot five seconds after you hit publish?

Engineer on the team here. We've spent the past months building the agent framework behind this launch, so let me share what we were actually going for: not another site scanner, but a QA engineer for your team.

Here's what that means in practice:

🤖 Agents that find problems before humans do. They sweep every page, desktop and mobile, in parallel, broken links, wrong phone numbers, stale pricing, brand guide violations and pin every finding on the live page. What used to take a team days happens in minutes, before a human ever opens the site.

🧠 Self-learning: from your checklist and your people. Reject a finding and it never comes back. Catch something the agents missed, and it becomes a check on every future run. But it goes further, the agents learn your team's review patterns, what your reviewers flag, what they let slide, what each client considers a dealbreaker. Week one it knows your checklist. Month three it reviews like your team does.

💬 Ask AI: your client knowledge, on tap. Every run builds up patterns about each client's sites: what breaks repeatedly, what's been flagged before, what that client cares about. Instead of digging through old reports, you just ask.

🔄 Two-way sync with the tools your team lives in. Slack, Monday, ClickUp, Asana, comments flow into your PM tool as tasks and data flows back. Close the task in ClickUp, it resolves in Superflow. No copy-pasting screenshots into tickets, no orphaned findings living in a separate dashboard nobody checks.

🔌 Fits the stack you already ship with. Webflow, WordPress, Shopify or a plain script tag. Your checklist, your platforms, your final say nothing ships until a person approves it.

The goal was simple: give every team the tireless QA engineer they could never afford to hire one that studies how your best reviewers work and keep the judgment calls exactly where they belong: with you.

Agents do the checking while you focus on the bigger picture.

An always-on QA teammate when you need one. 🔥

Watching multiple agents crawl a site in parallel and pin findings right on the page is amazing!!.

Agencies can now have QA Agent teammates who run 24/7 and never miss a beat!

Wow, we are preparing to launch this Sunday, and Superflow launched just in time for us to use it to QA our landing website and prod webapp. Amazing, super useful product.

All the best for your launch!!

Wow this is super interesting for me as a marketing lead. Since we are localizing our content, a lot of errors miss our fingertips.

Love the focus on automated website QA before launch day. Super useful build, congrats guys!

No more sweating over broken links or typos right before hitting publish. 😅 We built this to catch all those sneaky little bugs automatically, so you can hand off Webflow sites with total confidence.

Nothing about that agency's QA process was broken. It had just stopped fitting the number of pages. You framed that generously, and congrats on launch number seven.

I tried to scan my site on your website, but I got stuck on the install page. There isn't an option to install, and I'm not really sure what needs to be installed.

Hey Product Hunt 👋

I'm Prabha, one of the founders of Hubble.

If you're building something that needs a patient's records, there's no clean way to get them. A patient's history is scattered across every provider they've seen, and each system is its own island. Most APIs return the patient-portal subset rather than the full record, so teams either build per-provider integrations for months, fall back to faxes and manual portal logins, or push the problem onto the patient.

Hubble is the layer that goes and gets the rest. Your user verifies their identity once, and we assemble their records from across their providers and return them through one API. We connect into EHRs and HIEs where they work, and when an API comes back empty we keep going with browser agents that submit the request through provider portals and voice agents that call records departments on the patient's behalf. Every request runs on the patient's individual right of access, they review every source, and they can revoke at any time.

We're the team behind the AI agents at Grow Therapy and Amazon's One Medical, used by millions of patients. We kept hitting this wall ourselves, so we built the layer that gets past it.

Would love your feedback, especially from anyone who has fought with records retrieval before. I'll be here in the comments all day.

(PH exclusive: 7 days free. Reach out and mention this post, or email me at prabha@hubble.ai and I'll get you set up.)

Aaron here, co-founder of Hubble. The reason we built this: AI is going to transform healthcare. I launched Amazon One Medical's first production AI agent and found that the key limiting factor was context. Getting a patient's history meant tons of broken integrations with upstream systems like EHRs and payers. Prabha and I realized everyone (even ourselves!) was rebuilding the same broken pipe, every time. Hubble is that context layer built once for all builders who need access to patient records across domains (care delivery, insurance, consumer health, life sciences research, legal, etc).

Ask me anything about what we learned trying to do this the hard way first. 

Congrats on the launch

The problem you are solving is genuinely interesting especially with patient records being scattered across providers

Curious to see how Hubble handles the messy real world edge cases here and excited to see where this goes

@prabhadublish @aaronleon thx for making this. Sounds like it could be really powerful for builders. Curious about its application for caregivers. I’m a caregiver for my partner, building a a medical “bible” of sorts across notion and Dropbox of every visit, surgery, consult, rx, test, procedure etc…. Once Hubble locates and compiles a patients full record history and normalizes it, is it downloadable as readable files (PDF’s etc)?

Hey team! It sounds awesome. How do retrieve those records and why other APIs can't? Anyway, wish you all the best here!

This is awesome! It would be great it this paired with prehospital records (on an ambulance) because those interventions/vitals are often lost in the patient care journey

If you want to learn more about your company and use case, book time on my calendar here: https://calendly.com/prabha-hubble/30min?month=2026-08

One of your fallbacks is a voice agent phoning a records department on the patient's behalf. That call is the unglamorous part of this problem. Congrats on taking it on instead of leaving it to patients.

Awesome product - I was just switching primary care providers and had to stitch all my records together. How can I ensure that my records are safe with Hubble? Is it easy for me to remove any of the records I don't want Hubble to hold?

Which countries are supported? Only the US?

the voice agents calling records departments is the interesting part to me - what happens when a provider's line dumps into a callback queue or wants live verbal ID verification? does it hand off to a human or just keep retrying

the voice agents calling records departments is the part I'd want fail-tested. what happens when the person on the other end gets suspicious it's not a human and throws an unscripted question at it?

this is huge — going to different doctors always meant starting from scratch and they never had the full context. this is key to better patient outcomes 🔑

congrats on the launch!

Literally happened to me with Kaiser last week and it was a nightmare! Excited to see a product fixing these critical dependencies.

ElevenLabs MCP brings voice-agent management directly into Claude, so teams can manage their agents from the chat interface.

It reduces the need to switch between tools by letting users create, review, update, duplicate, and delete ElevenLabs agents in one place.

Key Features

• Create, update, duplicate, and delete agents.

• List agents and review their summaries.

• Check agent knowledge size, widgets, and links.

• Estimate LLM usage before making changes.

Benefits

• Simplifies voice-agent management.

• Makes agent configuration easier to review and maintain.

• Helps teams understand potential LLM usage before updates.

Who It’s For: Teams working in code, productivity, design, and sales and marketing.

Sign in to ElevenLabs through Claude to get started.

Prompt edits are the thing I'd iterate on most, and iterating means occasionally making an agent worse. Does a change made through chat get versioned on the ElevenLabs side so I can roll back to yesterday's prompt, or is the previous version gone once Claude writes it?

You kept review, prompt updates, duplication, and deletion in the same chat where the agents get made. Congrats on covering the tidy-up work and not only the building.

the "estimate LLM usage before making changes" bit is the one I'd want most - is that a token-diff estimate off the prompt edit or does it actually simulate a call against the new config?

if two teammates both have Claude open, what stops one person's duplicate/delete from stepping on an agent the other is mid-edit on? that's the part that'd worry me most about exposing this via chat

I use Claude to write prompts for ElevenLabs voice generation — is there already something that lets Claude generate the voice directly too, not just agents?

The usage estimate before making changes is a useful safety rail. For prompt or voice updates, does the MCP show a diff and require confirmation when projected usage rises or a handoff rule changes? That boundary would matter a lot for production voice agents.

Hi Product Hunt 👋

I built Shepherd because running multiple coding agents quickly became harder to manage than the code itself. Codex and Claude were spread across terminal windows, and I kept losing track of which agent was working, waiting for input, or already finished.

Shepherd is a persistent macOS workspace designed around coding agents:
- Run Codex and Claude across tabs, panes, and remote machines
- Keep terminal sessions alive even after closing the app
- Monitor every agent’s status in real time
- Let agents understand and control the current Shepherd workspace
- Open connected browser reviews from an agent and return your visual annotations as structured feedback
- Resume previous agent sessions and inspect their files, diffs, and Git history
- Connect to remote development machines over SSH and access their terminals, files, and local services
- Provide local STT model to vibe-code using your own voice

The macOS app is available today for Apple silicon Macs running macOS 14 or later. I’m also building an iOS companion for monitoring agents, receiving completion or attention alerts, give feedback to agent, control a terminal and reconnecting when away from the Mac.

I’d especially appreciate feedback on the agent-monitoring experience and whether Shepherd makes parallel agent work easier to follow.

This is Shepherd’s first public beta, so you may encounter bugs or rough edges. I’ll keep fixing them and exploring better ways to connect coding agents with Shepherd Terminal. Please try it, use it in your real workflow, and leave plenty of comments if you run into a problem or have an idea. Your feedback will directly shape what I build next. Thank you!

Thanks for trying it!

This feels built around how people actually use coding agents now.

I can already see this saving me from terminal window chaos.

the "sessions stay alive when the app closes" part is the thing I'd actually want, I've lost track of a long-running agent more than once because I closed the wrong terminal tab by accident. curious about the remote/SSH case though - if the connection drops mid-way through an agent actually writing or editing a file on the remote machine, does Shepherd just resume watching the same process, or is there a chance the agent's action gets cut off and you come back to a half-written file. that's the scenario that'd make or break trusting this for anything longer than a quick task

@kojunseo The part I'd want scoped is agents understanding and controlling the Shepherd session itself. If an agent in one pane can read and drive the workspace, can it see or act on a sibling agent's pane, or is each session's control surface limited to its own tab?

If you run into any issues, have questions, or need help getting started, please leave a comment here. I’ll be around all day and would love to hear your feedback!

@kojunseo Git history beside the agent context makes a lot of sense.

what did testing 200+ prompts teach you about product research?

I hunted AirBuddy 2.0 six years ago — and I've been following it's evolution ever since then. v3.0 is a massive upgrade and for anyone that uses Apple peripherals, you should really check this out.

Not only that — but I built a Raycast extension for AirBuddy so you can control all your devices from the comfort of your keyboard!

This is one of those apps you don't realize you need until you've used it. I especially like the battery pop-up and quick controls. Curious, was there a particular macOS limitation that was the hardest to work around while building AirBuddy?

Moving peripherals between Macs is the feature I have wanted for years. Right now I do the Bluetooth menu dance every time I switch from the laptop to the desktop, so if AirBuddy 3 handles a keyboard and mouse handoff cleanly that alone is worth it.

Moving peripherals between Macs without re-pairing is the feature I actually want - the usual Bluetooth handoff flow (unpair, forget, re-pair) is where I lose the most time switching machines. Is that peripheral move instant, or is there still a short handshake delay on the target Mac?

if i can later provide judgment as context for agents that will dope, the judgment layer should be coined as the new context layer :)

maybe an odd request, but I'd like to see some sample decisions to help jumpstart my activity on the app....the unknown size/shape of an opinion can feel daunting.

to answer your question directly - calibration absolutely makes sense outside forecasting, I just don't think most people would use that word for it. I screen early-stage startups fast, sometimes dozens a day, and the thing I never have is a record of how confident I actually was on a pass or a yes before I knew how it turned out. hindsight rewrites that instantly. the 30-60 day "would you still make this call" check-in is the part that sold me, that's the exact gap between a note app and a real record of judgment. one question - when the new info you tag as "positive" or "negative" mid-decision turns out to be irrelevant to the actual outcome, does that get factored into the calibration score at all, or does it just sit there as a log entry with no scoring consequence

Hey @rolandleth

Outcomes can be right for the wrong reasons, so tracking the reasoning & confidence feels way more useful than a normal decision journal.

Hey Product Hunt! 👋

2 years ago, we launched the first version of Clipwing here – a simple video editing tool that turns long videos into short clips.

Since then, Clipwing has grown into a video studio. We’ve made thousands of clips for startups, creators, and tech companies – and learned what really makes a clip worth posting.

But creating clips still takes too much work.

You need to find good moments, hire and manage editors, chase feedback across Slack and Google Drive, and move every clip through revisions and publishing.

So we built Clipwing Autopilot to keep the whole process in one place.

You upload one long video, choose your style, and get ready-to-post clips – without AI slop or the hassle of hiring and managing editors.

The clips are created using the same process we use in our studio:

🤖 AI can find moments using prompts tested on real client videos

🧠 our content team reviews and improves the clip ideas

🎬 a professional editor turns them into finished clips

Everything else also happens inside Clipwing:

📋 track clips on a board

💬 leave comments directly on the video

✅ review and approve edits

📅 schedule posts

♻️ turn one video into weeks of content

Our goal is simple: you keep recording, and Clipwing handles everything after that.

Would love to hear what you think 💙

LERA THE GOAT. Congrats on the launch Lera, Clipwing just keeps getting better

Cool! congrats

This is really cool, congrats on the launch. I am excited to see where this type of productized service and agency-ish work is going especially nowadays with AI. Well done!

i like the shift from product to service to the product again, its going to be a beast for creators by a creator

Congrats Lera! Two years from a small clipping tool to this. Good to see where it went 🔥

Congrats Isaac & team on the launch!

Finch sounds like a really thoughtful solution to a frustrating part of healthcare.
Curious what’s been the biggest challenge so far in getting patients to trust and use Finch?

What was the thinking behind the logo? Curious choice for a health app...!

Exciting!

Don't use Gauge, unless you want to win.

Hi Product Hunt! 👋 I’m Caren, co-founder & CEO of Agenda Hero. 

Today we're launching Dates. 

You can write together in Docs. Budget together in Sheets. Present together in Slides.

But where's the place for your Dates together?

I'm not talking about the Dates you go on. Or your individual personal calendar where you see your next meeting. This is about the Dates you share. Together. 

The school calendar you share with 500 other families. The theater schedule you share with the rest of the cast. Even the team offsite and sales kickoff agenda you share with your colleagues. Today these dates live in Docs, Sheets, and PDFs. Everyone has to reenter the same information as everyone else. Then it changes, and everyone’s calendar is wrong. 

Dates is a doc-simple way to make a plan for anything and share it with anyone. Bring your other calendars in: Google, TeamSnap, ParentSquare, ics feeds. Send the plan back out to whatever calendar each person already uses. Print it for the fridge, pull it up on a phone, and when the tournament moves, it moves on the team calendar, the family calendar, and Grandma's.

It runs on Agenda Hero Magic, so any text, image, or PDF becomes calendar events in seconds.

We’re incredibly grateful to everyone around the world who gave feedback and shared their magic wishes. You're the reason for the small details that bring big joy. A favorite one: bring in an ics feed, and you can hide anything on it. School calendar listing the 3rd grade choir concert when you have a 2nd grader? Poof. Hidden.

We shipped as early as we could, so expect some rough edges. Send us the bugs, and keep the requests, magic wishes, and encouragement coming. The very awesome Agenda Hero team will be in the comments all day.

One tip: try it on desktop. Mobile handles all viewing, but not all customization yet.

Free to start. Create one, or a dozen, at agendahero.com.

A school sends one PDF to 500 families. Why should 500 families have to rebuild the same calendar?
It sounds so absurd, but it's the way we've been doing it forever.

I feel so proud to be a part of the team that built this! We hope you all love it.

Once you experience the Magic of Agenda Hero, you’ll wonder why something like this hasn’t yet existed.

A small but mighty team that I’m so proud to be apart of! We built such a wonderful product and I hope it makes your life more magical!!

woow ! the pile of stuff I saved once and never saw again is enormous, and just knowing something finally tackles that gives me real hope

to answer the "what would you want it to do" question - I screen a lot of early-stage stuff and my saved links skew heavily toward pitch decks (usually a PDF link or a Google Slides/Docs link shared directly, not a public webpage) and X threads that are only visible if you're logged in and following the person. both of those are exactly the kind of content that a "fetch and read the page" approach struggles with, since there's either no public HTML to fetch or the content sits behind auth the extension isn't supposed to touch. is that kind of source on the roadmap, or is it intentionally staying scoped to publicly fetchable pages plus the four social sources you already support

The "browser bookmarks (automatic)" part is what caught my eye - since it indexes everything you bookmark by default, is there a way to exclude specific bookmarks or folders from ever being fetched/embedded? Some bookmarks are things like internal work docs or personal stuff you'd want searchable locally but not sent through any pipeline at all, opt-out per source doesn't quite cover that.

The number I'd watch there is the 90 seconds, not the 100ms.

I run a search index over about 10 million chess games, and the thing that caught me out was that whatever you extract at ingest quietly becomes the ceiling on what's findable forever after. When I improved the extractor, everything ingested before that point was still described the old way, and the fix isn't a deploy, it's re-running ingest across the whole corpus.

At 90s an item, a 10k library is roughly ten days of wall clock to redo. So the thing I'd want it to do that it may not: keep the fetched source and the generated description as separate layers, so a better describer can be replayed over pages you already have instead of going back out to the internet. I didn't split those at first and paid for it twice.

Indexing the frames rather than the caption is clearly the right call though. 'The reel with the one-pan pasta trick' only resolves if the trick is the thing that got indexed.

Wow! I have a teammate who absolutely needs this. Sending it to him now.

For me, I'd like the ability to query across LinkedIn and Facebook saved items too!

Hey hunters,

Said here, the builder behind CrewTower.

I built CrewTower to win back all the time I was losing while working with coding agents. We've all been there: running a few agents at once, then forgetting to reply to one and leaving it sitting there waiting on you for hours. By the end of the day you've burned both your time and a good chunk of your leftover tokens. That's exactly the problem CrewTower solves. That notch sitting right at eye level on your MacBook becomes your agent control panel. The second an agent needs a response from you, you'll catch it, review it from the notch, and take action without losing any time. It ships with sound notifications, a silent mode, swappable pixel art characters, plan reviews, the ability to jump straight into any session from the notch, or just reply from the notch without going anywhere at all.

I shipped it a week ahead of the Product Hunt launch, so CrewTower has already been serving its first users. I've been pushing out updates non-stop, big and small, to keep making it better. The feedback so far has gotten the product into a really solid place. Hope you enjoy using it.

No discount code this time, unfortunately. It already launched at early bird pricing, and I did everything I could to keep it affordable for everyone.

Got thoughts or suggestions? Drop a comment here or email support@crewtower.app.

Upvotes appreciated. Enjoy with crewtower.app!

"This is not an AI-generated comment. Fully written by me :)"

Hey @said_altan

Such a clever use of space.
Feels especially useful when you’re running multiple agents & don’t want to keep checking terminals

Hey @sidphoenix

I'm curious how much the graph structure improves agent performance compared to just throwing the same data into a vector store?

This hits a real agent problem: more context often creates more noise, not better decisions. Turning context into something agents can actually navigate feels much more scalable than endlessly expanding prompts.

Hi Product Hunt!

monolog started from a very personal frustration.

Whenever I tried to save a quick thought, I found myself thinking about titles, folders, tags, or where it should go. Then I realized I was already sending messages to myself in chat apps because it was much easier.

The problem was finding those messages later.

I actually tried to build this idea a few years ago, but stopped because I couldn't automate the organizing without giving that work back to the user.

AI changed that.

Now you can just write naturally. monolog understands schedules and tasks in the background, and semantic search helps you find old records even when you don't remember the exact words you used.

I built monolog mostly by myself, and I use it every day.

I'd love to hear what feels useful, confusing, or unnecessary after you try it.

Thanks for taking a look!

Hi everyone! 👋

I’m part of the team behind LayerProof, and I’m also the one running our social every single day, which means I’ve been living inside this product as its own toughest user.

So this launch is a big one for me 🎉

Being on both sides, building it and actually posting with it, shaped a lot of Matte 3.0.
If something annoyed me at 9 am while filling the content calendar, it usually became a fix.

For context on how far we’ve come: v1.0 could basically make you one image. v2.0 got smarter about captions and aspect ratios. Matte 3.0 is the version that actually changed how I run my week.

What I’m most excited to finally share:

• Brand kit: paste your brand once, and everything comes out on-brand. No more re-explaining colors, fonts, and voice to a tool every single time.

• Carousel storytelling: the big one. It builds a real narrative across images that follow Hook-Story-CTA structure. That used to eat hours of my week.

• Mark to edit: don’t like a line or a layout? Just mark it and refine it in place. No regenerating the whole thing and losing the parts you loved — you stay in control of the final result.

• Co-designed with 10 marketers across SaaS, FMCG, F&B, Consulting and more: we didn’t build this in a vacuum. You can feel their standards baked into the output.

• Ready-to-post: captions and formats done, straight from LayerProof. Idea to published without bouncing between five tools.

The mindset we kept coming back to: don’t just make one pretty image and call it a day; we want to actually grow your social media accounts — an on-brand feed, more followers, and high-converting content that really sells.

Would genuinely love your feedback. Drop your questions below, I’ll be here all day!

So proud to be part of this launch! 🎉

As a designer who uses LayerProof Matte 3 on daily basis, what stands out is how much it respects craft:

• Visual hierarchy: the Hook-Story-CTA structure isn’t just a template; it actually controls where the eye lands, how emphasis flows, and what each slide is doing. That intentional composition is the hardest part of carousel design to get right.

• Brand Kit as a real design system: not just “matches my colors,” but holds color, type, and tone the way a proper system should. That fidelity is usually what takes the most discipline to maintain by hand.

• Mark to edit: refine one line or element in place without the AI overwriting a layout you deliberately composed.
  
  A use case where it really shows is composing a multi-image carousel from scratch. I go from a blank frame to a hierarchy-driven, on-system set without fighting the tool for control of the details. Give it a try and let me know how the result lands!

How does pricing work? flat subscription, or credit-based per post/batch?

Hey Product Hunt 👋 Jordyn here, I'm part of the Marketing team at LayerProof!!!!

Thank you so much to everyone checking out LayerProof today.

Our own team was drowning in the "one message, too many platforms" problem every single launch cycle. Resizing was never the hard part, that's a solved problem. The real time sink was rewriting the same idea four different ways for LinkedIn, X, IG, TikTok...., and then still having to make sure everything stayed on-brand across all of it 😿 By the time all five versions were done, we barely had energy left for anything else on the to-do list, let alone the strategy work we actually wanted to be spending our time on. Entire afternoons gone just reformatting the same idea over and over 😩

So we built LayerProof to fix exactly that. Paste your brand once and every output stays on-brand, no more re-explaining your colors and voice to a tool every single time. Carousels now follow an actual hook, story, CTA structure instead of five slides that just repeat each other. And if one line or layout is off, you can mark it and fix it in place instead of regenerating the whole thing and losing the parts you liked.

We'd genuinely love for you to give LayerProof a spin today and tell us what you think! I'm dropping in all day to answer questions, so ask away, nothing is too small or too technical 🚀

What kind of hardware do you need to run this?

Hi Samuel, the part that stuck with me is that all this work no longer just vanishes when you close things. That really annoys me, so seeing someone treat it as worth keeping feels right, genuinely curious to see where this goes.