PH热榜 | 2026-08-02

Hey Product Hunt 👋

Maker here.

Most AI waits in a chat box. We built the opposite.

Zinley is your personal extension — own phone number, email, and computer. It answers calls, handles email, books stuff, follows up, and gets work done inside your rules. Remembers your people. Reports back in your language.

Not another chatbot. We extend you so the people around you can actually reach “you” when you’re busy.

Curious what you’d hand off first — and what would make you trust an AI with your number and inbox.

Here all day. Roast us.

— Khoi + team

The "people call it, not your personal line" framing is a good trust argument, but it cuts both ways. When Zinley answers a call, does the person on the other end get told upfront they're talking to an AI, or does it just present itself as reachable the way you would be? A few US states already require that disclosure for AI phone agents, worth having a clear answer to that before launch rather than after.

Giving it its own phone number instead of just inbox access is the part that'll get people to actually trust it with something.

the caller-ID angle is what worries me more than prompt injection - if part of how Zinley decides "is this a known relationship" leans on caller ID or the from-address, both are trivially spoofable. someone who's done a bit of recon on you (knows your assistant's name, knows a real contact's number) could call in pretending to be that trusted contact and get bumped into the more autonomous tier before Zinley ever has to guess anything risky. is there any out-of-band verification for "this really is the person Zinley thinks it is," or does trust level still ultimately trace back to caller ID/email address matching a stored contact?

I was wondering about its boundaries, how does Zinley decide when to handle a request itself versus asking for your approval first?

Ah! I like the idea of giving AI its own phone number instead of keeping it trapped inside another chat window. :P

Congrats on the launch!

Congrats on the launch! Can users inspect, edit, export, or fully delete Zinley’s memory about their contacts and relationships? That control feels essential for something this personal.

Rare to see an AI tool that discloses itself upfront instead of trying to pass as human. The escalate-to-human-on-uncertainty design is the right call, most agent products optimize for autonomy first and trust second. Curious how you handle the handoff mid-call when something crosses that uncertainty line, does the caller notice a pause, or is it seamless?

the own-phone-number angle is the interesting part, most AI assistants die the moment someone outside your chat needs to reach them. curious how it handles calls it shouldn't answer for you though, like how granular the rules are before it hands something back to the human

It would be awesome if Zinley could recognize recurring requests and proactively automate similar tasks.

@jasonzinley Congratulations. And happy product launch.

Congrats @jasonzinley @peter_bakas How does Zinley adapt its tone when talking to family, customers, or business partners?

Congratulations on launching. I have some use cases in mind altogether.

What were some use cases that your initial users used it for that surprised you??

This caught my attention. We live in times of dramatic distractions. Products that reduce communication overload instead of adding another inbox will always find an audience.

Busy professionals would rely on something like this every single day once trust is established. Congrats on shipping this. Good stuff.

Can Zinley negotiate meeting times or bookings completely on its own within predefined limits?

Trusting an AI with email feel natural but handling over a phone number feels like ultimate test . I'd start with filtering out spam/cold calls before letting it talk to actual clients.

Also how it handles edge cases when someone speaks off-script on the call ?

What happens if someone asks Zinley to make a decision that's outside the rules you've defined?

The detail that grabbed me is that it has its own phone number and is reachable by the people around you, not just by the account owner. I build voice AI that places daily outbound calls to aging parents, and the trust barrier on an unknown number is the hardest part. How are you handling that first-contact moment so people actually pick up instead of treating it as spam? And does the memory of "your people and relationships" persist across calls, or reset per session?

An assistant with its own phone number is a different category from one sitting in a chat window. Most of these are built so only I can talk to them, this one the people around me can reach.

Congrats on the launch. Is there an analytics dashboard showing which tasks Zinley saved you the most time on each week? That will help determine true value of the product.

It's good to see the focus is on relationships and memory instead of just automating isolated tasks. Congrats!

MIT-licensing this after building it solo for your own workflow is a solid move, way better than watching people pay $29/month for something this scoped.

love seeing more open source tools like this. the fact that everything runs locally with no account or subscription is a huge plus. Congrats on the launch!

The on-device captions are the thing I'd actually use here, the cloud ones get awkward when the demo shows unreleased screens. Does the caption layout follow the export frame though? I crop demos down to 9:16 a lot and burned text is always the first thing that gets chopped.

Love that captions are burned on-device, keeping demos private while still looking polished is a smart touch.

Congrats on the launch!

I switched from Screen Studio to Recordly. What sets this apart from Recordly?

Loving the app so far! I did run into a bug on Windows 11 that might help.

When I try to export, the process fails with:

EncodingError: Decoding error.

Also, the custom background I selected isn't applied during export—it ends up rendering as a solid black background instead.

Not sure if the two issues are related, but hopefully this helps. Keep up the awesome work!

Pretty nice app, I'll take a look. One piece of advice, though: having it signed is very important, especially for MDM-managed macOS computers in companies. Unsigned commands for Gatekeeper can be blocked for users.

Congratulations on the launch! I've been looking into video editors recently and Capptivo looks like it fills a nice place to help with demos.

Do you have any plans to add presets for recording iOS / Android device demos? I'm a mobile engineer and one of the problems I have is making sure app previews have the right setup when uploading them to the store. That would be a great addition to see.

Congratulations again! 👏

Glad to see more options! Hopefully can cancel my Screen Studio

Good timing — I just skipped the video/demo field on my own Product Hunt draft because I didn't have a clean way to record one without a subscription tool. The click-based auto-zoom and local processing (no cloud upload for an unreleased product) are exactly what I'd want for a quick product demo. Does the caption burning work well with narration you record separately, or is it built around live mic capture during the recording itself?

This seems to be a bug. After selecting the recording area, the application became completely unresponsive—I couldn't start the recording or even close the window. I had to restart my computer twice to resolve the issue.

this is amazing dude! gonna use it for all my product launches going forward :)

I tried with my Mac. macOS 26.5.2. After donloading, I try to open and:
"Capptivo.app" Not Opened

Apple could not verify

"Capptivo.app" is free of malware that may harm your Mac or compromise your privacy."

I know Apple is very strict, perhaps too much? but I think it is important to polish this security issues first. We, the users, cannot know all those details, what is right and what is wrong and in case of doubt we trust Apple first. This is what we know is always in our side.

respect for open-sourcing this instead of just underpricing Screen Studio. the codesigning thread caught my eye though - Apple Developer Program is $99/year out of your own pocket for a free tool with no revenue, and that's a recurring cost, not a one-time fix. have you thought about how that gets sustained once it's not just you personally eating the bill, e.g. sponsorship/OpenCollective, or are you treating it as a cost of doing business for now? cross-platform + free + signed is a great combo for adoption but the signing part is the one piece that doesn't stay free by itself.

on-device captions plus no signup wall on export is a combo that basically doesn't exist in this category right now.

How does it compare to @Velo?

👋🏻 Hey Product Hunt, I'm Andras, founder and CEO of YourSitee.

YourSitee is a visual link-in-bio platform for creators, businesses, communities, and anyone who needs one clean place for their online presence.

We built it because a lot of bio pages still feel like either a plain list of links or an overbuilt creator stack. We wanted something in the middle: simple to set up, flexible enough to feel personal, and useful enough to show what people actually click.

With YourSitee, you can:

🛠️ Build a customizable page for your links, socials, images, events, and content

🧩 Add 20+ widgets, from videos and posts to markdown, countdowns, events, and custom sections

📊 Track visits, clicks, countries, and top-performing links in one simple dashboard

🆓 Keep the core experience free, and try Pro for 30 days!

✨ Use the AI Editor to import your existing Linktree page or social profiles by pasting a URL

If you already use Linktree, you can paste your Linktree URL and YourSitee will help you create a starting version of your page for free. Migrating should not be the reason someone stays with a tool they no longer love.

Try it here: https://yoursit.ee

The most useful thing you can do today is build or import a page and tell us what feels good, what feels weak, and what would make you switch from your current setup.

@abelherfort and @szalovszky, my co-founders and close friends, are here too. They have started two focused threads ↓ below: one for a founder and investor roast, and one where you can share your page and get direct feedback.

If you want the longer story behind our launch, we wrote it here.

One small note: if you have 10k+ followers on any platform and currently use a website builder, Linktree, Beacons, or another link-in-bio tool, we may have a custom offer for you. Email me with your link at: andras@yoursit.ee

My question for the PH community:

What makes you stay and explore someone's personal page, and what makes you leave immediately?

I'll be here all day answering questions about our journey so far, and we'll award badges to some of the most thoughtful, helpful, and interesting comments throughout the day.

Thanks for taking a look!

— Andras

Hey Product Hunt!

I'm Abel, co-founder and CTO of YourSitee. I work on our backend architecture, integrations, and the systems that keep profiles fast and reliable as the product grows. Outside YourSitee, I'm a platform and backend engineer at BlackRock.

I'll be here for any technical questions.

Founder and investor roast thread 👇🧵

What would need to be true for YourSitee to become a durable company, not just a useful product? Roast our positioning, defensibility, business model, or anything else. Honest criticism is more useful than polite support.

Nice launch @andrasczeizel qq Is there an option to password-protect or hide specific widgets/links behind an email capture gate for digital downloads?

hey Product Hunt, David here. i'm a co-founder and lead engineer at YourSitee.

i work mostly on the interface and product experience, plus parts of our service architecture. the hard part has been keeping the editor simple while giving people enough freedom to make pages that don't all look the same.

👇 sitee showcase thread 👇

if you build or import a page today, drop the link below. i'll try to check out every page shared here and reply with one thing that works well and one thing i'd improve.

we'll also choose some of our favorite profiles from the thread to receive 6 months of YourSitee Pro for free.

The AI Linktree import is genuinely smart, saves so much tedious copy-pasting. Love how the widgets feel curated instead of bloated, gives the whole page a clean editorial vibe.

I actually really needed something like this since linktree feels so outdated on the free plan!

What kind of integrations do you have?

been a user for a while it was still in beta and so much small adjustments really made this one shine!

thanks team for pushing the details and congrats

Pay for what you run sounds obviously fairer than a subscription, and it has one nasty failure mode: people start rationing. The moment someone can feel the meter, they stop asking the second and third question, and those are usually the ones that get them the answer. Watch actions per session in week two against week one. If that drops, the pricing is teaching people to use it less, and "under a dollar for an afternoon" won't save you.

"A black window that expects you to already know the words" is exactly why terminal tools stay niche no matter how good they are.

Interesting angle — the terminal is doing more work than people realize, and it's not the typing.

When Claude Code generates something that doesn't compile, the terminal is where the loop closes: the agent sees the error, reads it, and retries. Take the terminal away and someone still has to close that loop. Does Lumichats surface the failure to the user, or does it self-correct silently before they ever see it?

I ask because in my experience that's where the whole "no-terminal" promise usually breaks: the moment a non-technical user is shown a stack trace, you've lost them — but if you hide it and retry blindly, you burn tokens and they wait without knowing why. Curious how you handled it.

The permission split (ask / auto-apply / read-only) covers whether it asks. The part I can't picture is what the ask looks like for someone who doesn't know the words. I run a terminal agent every day, and an approval prompt is only a safety feature because I can evaluate it — "run this command, write those files" carries information for me, so my yes/no does too. For the researcher with 200 PDFs, the same prompt is noise. By the third one they're clicking yes on reflex, and the ask has quietly become a ritual instead of a decision.

Do you translate each action into plain-language intent before asking — "unpack the images from this PDF" rather than the command itself? And if so, how do you handle the gap when the description and the actual action drift, since the user is approving the sentence, not the command? You solved the error side with "demoted, not hidden" — one line, expandable to the full trace. Wondering if approvals get the same treatment, because that gap between what's said and what runs seems like the hardest part of "no terminal" to me.

Zero data collection and fully offline is the first thing I check before recommending a tool on a security-sensitive project. That alone covers a lot of ground. What I want to understand is the model isolation side: does the app sandbox the model process away from the filesystem, or does it get direct read/write access to the project directory? And for the local models, are they bundled or bring-your-own?

the ask-before-changes / auto-apply / read-only split covers permission, but what about undo? if it's written a real .docx or edited files in a folder and I don't like what it did, is there any versioning or rollback, or is the expectation that I'm backing my own files up before pointing it at a folder? that's the part that'd decide whether I trust it on anything important versus a scratch directory.

The CPU only inference is the quietly underappreciated constraint here most offline AI projects quietly assume you have a CUDA capable GPU, so running on any modern laptop without that dependency

What's the difference with Claude Cowork ?

Any plans to launch the iOS version too?

honestly love that the shortcuts work system-wide without sending audio anywhere, that on-device approach feels rare these days. nice work on keeping the core stuff local.

Love that the transcript search lives right inside the app instead of needing yet another tab, and the on-device approach means my random shower thoughts stay mine. Clean execution.

I got hit with a bunch of Wispr Flow promos this morning and clicked on one for later viewing when I saw your launch. My first question was why another and tough timing! But reading it's local changes this (assuming Wisp Flow is not). I'd rather have yours! Good luck today.

the local-first angle is what would actually get me to try this over Wispr Flow or Superwhisper. I dictate a lot of technical stuff - API names, package names, product names that don't exist in any dictionary - and that's usually where on-device models fall apart compared to cloud Whisper. does Zen Whisper let you teach it a custom vocabulary/glossary of recurring terms, or is accuracy on that kind of jargon just whatever the local model ships with out of the box?

There are one or more voice dictation apps everyday on PH. What makes this different?

Zen Whisper is live on Product Hunt today. 🎉

We built it because Mac dictation still felt too limited, too cloud-first, or too awkward for real everyday work.

Zen Whisper is our take on a better workflow:

• 🎙️ hold a shortcut and dictate into the apps you already use

• 🔒 keep core speech recognition local on your Mac

• 🌍 work across 100+ languages, with especially strong multilingual workflows for India

• 🎧 transcribe files, media, and supported links

• ✍️ use snippets, transcript history, and writing tools in one place

It starts with a free trial, stays usable on the free tier after that, and scales up if you need more.

We’d love your feedback, especially on:

• dictation quality

• multilingual support

• latency and responsiveness

• anything that still feels clunky in the workflow

Happy to answer every question here today. 🙏

On-device is the right call here — I've tried a couple cloud dictation tools and the round-trip latency makes them unusable for anything longer than a quick note. Does accuracy hold up for technical vocabulary / code-adjacent terms, or is it tuned more for plain prose?

Really like this idea! How intensive is this on computer memory?

Zen Whisper is live on Product Hunt today. 🎉

We built it because Mac dictation still felt too limited, too cloud-first, or too awkward for real everyday work.

Zen Whisper is our take on a better workflow:

• 🎙️ hold a shortcut and dictate into the apps you already use

• 🔒 keep core speech recognition local on your Mac

• 🌍 work across 100+ languages, with especially strong multilingual workflows for India

• 🎧 transcribe files, media, and supported links

• ✍️ use snippets, transcript history, and writing tools in one place

It starts with a free trial, stays usable on the free tier after that, and scales up if you need more.

We’d love your feedback, especially on:

• dictation quality

• multilingual support

• latency and responsiveness

• anything that still feels clunky in the workflow

Happy to answer every question here today. 🙏

the "type a note for a week then stop" problem is real and voice is the right fix for it. what I'm curious about is the correction flow, since parsing is never going to be 100%. if I say "thirty five for lunch" and it hears "thirty five fifty" and files it under the wrong category, how many taps does it take to fix that transaction after the fact? that gap between "voice capture is fast" and "fixing a misparse is annoying" is usually where habits like this quietly die a second time.

Love it Santiago! Having spent control is key and you're making it so easy! Love to see this kind of projects launching here and wish you all the best

Voice capture is the one thing that keeps me tracking expenses for longer than two weeks. Do categories come predefined or is it a blank slate? And I would love to know if there is a monthly summary that shows spending patterns over time, because a raw transaction list is where most finance apps lose me after the first week.

Sorry I don't get it. What is it? Speech to text with database of what I spent?

Storing API keys in Credential Manager is the right call. One thing I would want to know: does each model profile get its own isolated key, or is it one key shared across all agent sessions at once? Running a client project next to a personal one, I would want those credentials fully separated at the profile level.

finally something that treats windows like a first class citizen for agent workflows. the grid layout for multiple ptys is honestly the part that sold me, way easier to keep track of what each agent is doing

the recoverable part is what caught my eye more than the grid layout. when you say it resumes native sessions, does that survive an actual Windows update reboot or a crash, where the PTY process itself is gone, or is it more like restoring the pane arrangement and scrollback while the underlying agent process has to be kicked off fresh? that distinction matters a lot if you're mid-task on something long running.

the approval notification thing is honestly such a thoughtful touch, like that's exactly the kind of thing where you're away from your machine and an agent is just sitting there waiting on you. nice execution.

Alerting when an agent needs attention is the right primitive. The one nobody builds is the opposite alert, for the agent that's quietly burning your tokens on a loop going nowhere. The profiles already hold the API key, so you could put spend per session next to each PTY tab, and that number changes behaviour faster than any layout feature will. Running two agents at once and having no idea which one ate the budget is the part that actually costs me.

Windows-native is the underserved half of this space — most of the agent tooling assumes macOS, and the WSL workaround means your agent and your editor disagree about what a path is.

Does it run the agent inside WSL, on native Windows, or does it let you pick? I ask because the path translation is where I'd expect the friction to actually live. An agent that writes /home/user/... into a config a Windows tool has to read is the kind of thing that works in the demo and breaks in week two.

Storing API keys in Windows Credential Manager instead of a config file is the kind of detail that gets this trusted fast.

Your answer to Omri is the most interesting thing on this page and nobody has followed it.

Files already written remain on disk, and the in flight operation does not survive. So after a crash the filesystem has moved but the agent's belief about what it did is gone. On resume it reads a conversation that stops mid action, and nothing in that transcript tells it whether the tool actually completed. It cannot separate "the tool ran and I never saw the result" from "the tool never ran". If the action was idempotent that is harmless. If it was a migration, a push, a POST or an append, then redoing it and skipping it are both wrong and it has no basis to choose between them.

The useful part is that Termexo is the only component that knows a crash happened. The CLI does not, because from its side a resume looks like a resume.

Which means the fix is not process checkpointing, it is a note. On a resume that follows an abnormal exit, inject something the agent will read: this session was interrupted at 14:32 partway through an operation, verify state before continuing. That turns a confident silent continuation into a cautious one and it costs you one string.

At the moment a clean resume and a resume after the machine died look identical to the agent, and those are the two cases that most need to look different.

Asking as someone who runs agents on Windows daily: does the app currently know the difference between being closed properly and being killed?

The "notified when an agent needs approval" bit is the killer feature — babysitting permission prompts across parallel sessions is exactly what breaks the multi-agent workflow right now. How are you detecting the approval state, hooking into the agent's own permission protocol or parsing the PTY output? And does session resume survive a full machine reboot or just app restarts?

The bring-your-own STT/TTS/LLM flexibility is what stood out to me, since most voice platforms lock you into one stack. I've been building voice AI for eldercare in the US, and latency plus natural turn-taking were the two hardest things to get right. I'm curious how you handle code-switching mid-sentence, where a caller mixes English and a regional language freely. How does Bolcho detect the language boundary in real time?

This is actually a really good idea.
Just wanted to know whether you are considering dialects as well. Often, I have seen voice agents though they have Indian language support, they misinterpret grammar or dialect

latency and code switching are both taken above, so the one nobody has raised: in a support call the hard part is not the sentence, it is the identifier inside it.

order number, phone number, pincode, an email spelled out. a language model cleans up grammar, it cannot repair a misheard digit, and that failure is silent. you read back the wrong order and the call still sounds perfectly fluent.

confidence scores do not help either, because seven and nine are both confident. what helps is validity rather than confidence, checking the captured string against the real list of orders before the agent says anything back. that turns a recognition problem into a lookup, which is a much easier thing to be right about.

genuinely asking on the multilingual side, since you have the data and i do not. do your callers say digits in the same language as the rest of the sentence, or do numbers come out in english inside a regional sentence? if it is the second, the digit is crossing a language boundary in the middle of the one field you least want wrong.

does bolcho let an agent validate a captured field against the business's own data mid call, or is capture whatever the stt returned?

Does it cover all Indian languages?

Not locking people into one LLM/telephony stack is smart, that's usually the first thing devs want to swap out once they hit real usage.

Building on Artur, from having built a small invoicing tool myself. The thing I would want to know before trusting this with real clients is what happens to an invoice after it has been sent.

Notion is mutable and a PDF is a snapshot. If somebody edits or deletes a time entry after the invoice went out, the document your client is holding and the data in Notion now disagree, and nothing reconciles them. Worse, if the invoice recomputes from live sessions, then hitting Load and Generate again next month can produce a different PDF carrying the same invoice number. That is the one thing an invoice must never do, because it is a commercial document rather than a report.

So the question is simply: are line items frozen at the moment of generation, or recomputed every time?

Two related ones that bite everybody who builds this.

Does a session get marked as billed once it lands on an invoice? If Load pulls unbilled time and nothing writes back, the same hours can end up on two invoices, and you find that out when a client notices rather than when you do.

And where does the invoice number come from? If it is a Notion property then somebody can duplicate a row or edit the field, and you get gaps or duplicates. For UK VAT the sequence has to be unique and unbroken, so it wants to be issued by something a human cannot type into.

None of this matters until the month a client queries an invoice. Then it is the only thing that matters.

ooh this is so cool!! does it track your screentimes and stuff so you can show proof?

That's time saving, but can the invoice be customized with the markings of the business?

Hey Product Hunt!

I built my own tool to solve an annoying problem. And it works!
It's called FreqWave EQ and it brings the audio control I need straight to my browser.

Key features:
- 8-band parametric equalizer
- Voice Enhancer - four modes for speech and dialogue
- Real-time spectrum visualizer

It’s completely free on the Chrome Web Store.
I'd love to get your feedback and feature suggestions!

nice, this is the kind of thing that should've been built into Chrome years ago. question on the day-to-day workflow: does it remember a preset per-site automatically (podcast site gets the voice mode, music site gets flat/whatever I set once), or is it a global setting I'd be re-adjusting every time I switch between a podcast tab and a music tab? that persistence detail is usually what decides whether I actually keep using an EQ extension after the first week or forget it exists.

Real-time EQ in the browser runs into the same wall every extension in this space hits: sites using EME for protected playback block extension access to the audio graph entirely.

Which tabs does it actually work in? I'd want to know the boundary before installing rather than discovering it on the one site I care about. Since it's open source, is that documented anywhere, or is it more "try it and see"?

Actually tried it on a podcast that always sounds like it’s recorded in a tin can, and the difference was pretty immediate. The compression mode is a nice touch too, kind of wish more browsers had this built in.

The frustrating part of payments is the thing you have removed, so the useful question is where it went rather than whether it is gone.

Onboarding and KYC take days because somebody has to be accountable for who is collecting money from the public. That requirement does not disappear when the integration gets faster, it moves. So which is it: are you the merchant of record with builders sitting underneath you as sub-merchants, or is the KYC deferred rather than removed? Both are legitimate and they fail very differently. If it is deferred, the moment it lands matters enormously, because finding out you cannot withdraw after taking real money from real customers is a much worse problem than a slow start.

The related one is termination. If you are the merchant of record then my ability to keep collecting depends on your risk appetite, not mine. What does notice look like, and what happens to funds in flight? Nobody asks that until it has already happened to them.

Second thing, and it comes from this being a skill rather than an API. A payment link is a request for money in my name. If an agent can generate and send one, the failure is not a bad log line, it is a wrong amount, a wrong currency, or a description a customer reads on their statement and does not recognise. And it is quiet, because they either pay or they do not, and neither outcome tells me something went wrong. Is there a point where a human sees the amount before a link leaves, or is that entirely left to whoever installs the skill?

The setup lag is what kills the 'just try it' window. We run experiments adding paywalls to new features, and merchant onboarding turns that into a multi-week detour before we even know if users will pay. One question on the global side: when you say 'accept global payment methods,' does that include the payout leg, or mainly acceptance? Outside the US that's usually the harder half.

A blank textarea is where customer testimonials go to die. 💀

When you ask a happy customer to write a review, you aren't just asking for feedback—you're handing them instant writer's block.

PraiseEngine completely rewrites how B2B teams collect, curate, and deploy social proof:

1️⃣ The 3-Question Interview: Replaces intimidating open-ended textareas with a 30-second conversational Q&A that guides customers through their real results.

2️⃣ Grounded AI: Extracts stated facts and outcomes—no hallucinated fluff—and structures them into a compelling narrative with an Approve-First moderation gate.

3️⃣ The SEO Flywheel: Turns static testimonials into programmatic backlink engines using indexable public profiles and native Review JSON-LD schema.

Hey Product Hunt! 👋

I’m Austin, founder of @Salestrics. Today we’re launching PraiseEngine.

The Problem:
A blank textarea is a terrible customer interview.

When you ask happy customers to "leave a review," they hit writer's block. You get generic one-liners, or worse—nothing at all. On top of that, auto-publishing reviews creates risk, while manual collection leaves proof scattered across Google Docs and Slack messages.

How PraiseEngine Solves It:
We built an AI-native review loop that makes customer feedback frictionless and accurate:

- 3-Question Adaptive Interviews: Instead of an empty text box, PraiseEngine asks three short, context-aware questions tailored to your business.

- Grounded AI Drafts: It turns customer answers into a polished, editable review—using only their actual words (zero AI hallucinations).

- Approve-First Moderation: You stay in complete control. Approve or reject before anything goes live.

- Dual Embeds & SEO Flywheel: Drop a top 3–5 review widget on your site, or send buyers to your fully indexed public profile powered by Review JSON-LD schema and backlinks.

PH Exclusive Offer 🎁
Every plan comes with a 14-day free trial (no credit card required), but for the Product Hunt community, we're opening up all 3 plans FREE for 3 months so you can test out custom-branded widgets and white-label options. Use promo code PH2026 after your trial to claim.

We’d love for you to test out the adaptive interview right on our homepage and let us know your thoughts in the comments below! What’s your biggest friction point with collecting customer proof right now?

Launch Day Changelog:

Pricing & plans

• Free plan, forever — New accounts start on a permanent Free plan with no credit card. The old 14-day trial is gone; anyone still on a trial moves to Free when it ends.

• Clearer paid tiers — Starter ($29), Growth ($79), and Agency ($199) each have defined team seat limits: 1, 5, or unlimited teammates.

• Your brand on paid plans — All paid plans remove the “Powered by PraiseEngine” badge on widgets.

Branding

• Custom profile colors — Paid plans can set accent, background, and text colors on public review profiles (/reviews/your-company).

• Custom widget colors — Same paid-plan access for embed styling (already available; now aligned with profile branding).

Widgets

• Widget language — Choose a language when creating collection or display widgets. The embed UI, success messages, and AI interview questions adapt to that language (11 locales supported).

• Four display layouts — Display widgets now support Standard, Grid, Carousel, and Marquee. Pick a style when creating or editing a widget.

• Sortable review widgets — Choose how reviews appear in display widgets: highest rated, lowest rated, newest, or oldest.

• Delete widgets — Remove display or collection widgets you no longer need from the dashboard.

Reviews

• Delete reviews — Permanently remove any review from your dashboard (with confirmation).

• Highest rated first, by default — Reviews now sort by rating everywhere: dashboard, public profile, and widgets.

• Filter and sort in dashboard — Filter reviews by status or website, and sort by rating or date.

• Sort on public profiles — Visitors can change how reviews are ordered on your public review page.

Marketing & help text

• Updated copy across the product — Landing page, pricing, dashboard help text, and onboarding now reflect layouts, sorting, delete, and the Free plan.

By request, we now have a language selector built into the widgets.

👋 Hi Product Hunt! I'm Maulik, founder of @Apoointly

When we started speaking with healthcare practices, one problem kept coming up: missed patient calls.

Front desk teams are constantly balancing phone calls, walk-in patients, scheduling, and administrative work. No matter how dedicated the staff is, it's simply impossible to answer every call.

That's why we built @Apoointly.

Apoointly is an AI Front Desk for healthcare practices that answers calls, schedules appointments, answers common patient questions, and supports clinic staff 24/7.

Our goal has always been simple: reduce administrative workload so healthcare teams can spend more time with patients.

Whether it's after-hours calls, appointment scheduling, or repetitive patient questions, your front desk shouldn't have to do it all alone.

We're just getting started, and we'd love to build this together with your feedback.

Thanks for checking out Apoointly. We'd love your feedback and will be here all day to answer your questions.

What kind of integrations do you have?

honestly the move of putting everything in whatsapp instead of forcing another app download is pretty smart — feels like you actually thought through how people already live on their phones.

@itsbhavesh Thats a great idea and very affordable. With WhatsApp reminders, it's easy for everyone to understand and take the medicine on time. The before and after lunch is a nice touch I see that you have a pharmacy partner option as well. Kudus on the launch.

Hi everyone, I’m Vignesh, founder of KlientFlow 👋

I built KlientFlow after struggling to track outreach for my editing agency. Excel and Notion became messy, while traditional CRMs felt overwhelming. The turning point came when a prospect said, “Get back to me in three months.” I realized a reminder wasn’t enough—I also needed to remember the person and the context behind our conversation.

KlientFlow automatically schedules follow-ups, helps preserve client context, prepares editable AI-assisted drafts, and opens the right LinkedIn, Instagram, X, or email conversation when it’s time to reply.

This is Beta v0.5, and we’re still improving it. I’d love your honest feedback, especially if you’re a freelancer, solopreneur, or founder who relies on outreach. Feel free to connect with me on X. Cheers! 🚀

Hey, I like the idea but wasn't able to try it out due to an issue with Sign in with Google!

Local processing for image merging is the version I'd trust, since the files people batch together are usually the ones they'd rather not upload — scans, screenshots with account details in them, documents.

Where does the 30 limit come from? Browser memory, or a deliberate cap to keep it responsive? I ask because 30 is oddly specific, and if it's a memory ceiling then image size probably matters more than count — 30 phone photos and 30 scans are very different asks.

Hey hunters, maker here.

I kept seeing the same gap at company after company. Hundreds of happy customers, but only a handful of case studies. Not because the stories aren't there but producing them is brutal. Chase the customer, schedule a 45-minute call, transcribe it, stare down a blank doc, write, edit, chase approvals. Three to six weeks per story.

The thing that unlocked it for me was that almost nobody fills in a written testimonial form or answer questions for 45 mins, but almost everybody will happily talk for 5 minutes async with an AI.

So StoryVoice works like this: you send your customer a link. They talk to our voice AI for about 5 minutes in an engaging conversation, whenever suits them. StoryVoice turns that conversation into a high-quality publish-ready case study in their own voice, plus a LinkedIn post, an X post, a pull quote, and a sales email.

Two things I care about that I'd love you to poke at. It computes headline metrics only from numbers your customer actually said, and if an interview doesn't contain a real story, it declines to write one instead of faking it.

The goal is simple: a case study should cost 5 minutes of your customer's time, not weeks of yours.

You can try the AI interview yourself at storyvoice.io/demo (no signup), and there's 20% off any plan for launch week. Would love your feedback. Happy to answer anything.

— Venkat

the "decline to write one instead of faking it" bit is the detail that'd actually make me trust the output. what happens on the customer's end when that happens though - do they know their interview didn't produce anything, or does it just quietly vanish from the maker's dashboard? if a customer took the 5 minutes and nothing came of it with no explanation, that seems like the one way this could burn the exact goodwill it's built on.

Said I'd come read this properly, so here it is. The five minutes isn't where our case studies died, approvals were. Every one I've tried to ship got stuck at the customer's manager or their legal person, so a fast draft turns a six week problem into a five week one. The thing I'd pay for is the approval step: send them the piece with their own quotes highlighted and a single approve link, so nobody has to read a whole doc to say yes. Does the customer see the draft before the maker does?

Hi guys,

I'm the solo dev behind my take on a transaction tracker with a silly mascot sprinkled on top. I wanted an easy fun way to sort and view where my spending is month over month without the hassle of managing spreadsheets, exporting bank docs, and having to constantly update my tracker.

Over the last 5-6 months I built exactly what I wanted: my app Taffy pulls your transactions, stacks them like a deck, and groups all similar transactions for you to sort into a bucket with a single tap. Repeat transactions are automatically sorted into that bucket with options to split categories or edit the amount (for those split bills). Then it shows you where it all went, your trends over time, and your top merchants.

Connections run through Plaid, so it never sees your bank login. Works with Canadian and US banks, iOS only. Free with one bank connected — the whole app, every feature — and connecting more than one bank is where I'll charge a monthly subscription.

This is genuinely a "does anyone else want this?" moment for me, so I'd love your honest take — what's confusing, what's missing, what you like/don't like.

App store link

Full demo for reference.

Finally, something that just shows me where my money went without making me plan six months ahead. Tapping each purchase into a bucket felt almost fun, like sorting receipts but quicker.

Cyclops roasting me for losing it over a parking ticket then reading the calm version in an Irish accent genuinely cracked me up. The petty mode is dangerous but the talk-me-out option is unexpectedly useful.

The roast is the interesting design choice. A tool that just says "maybe don't send that" is a scold, and people uninstall scolds. Making it funny means the intervention costs the user less pride than it otherwise would.

Does the humor stay funny on the tenth time, though? That's the thing I'd expect to decide retention. A joke you've heard before while you're genuinely angry might land worse than no joke at all.

Small over-the-air (ota) updates prelaunch. Good thing about ota updates are that the user doesn't need to download a new app version, it just silently delivers.

Tweaked the personality, range and accents of our resident mascot, Cyclops, who also got his own backstory for why he has so many accents and knows so many languages.

Hey yeah that was an early UX worry and I've tried to make it vary a lot. Maybe it's not a daily use app anyway and possibly a party truck. But I will work on that algo for making it more varied.