PH热榜 | 2026-07-11

me thinks the effect collection looks useful for many apps? could lightweight quality settings help developers support more devices easily?

@maxim_troshin Love that everything runs client side. Huge win for privacy and latency.

@anton_tushmintsev How do you decide when adding another effect is actually worth it?

@maxim_troshin Have you had anyone choose a simpler solution instead of this and why?

@maxim_troshin Was there ever a point where you felt the SDK was becoming too feature heavy?

@maxim_troshin Do most developers actually use all these effects or just a handful?

@maxim_troshin Which feature has delivered the biggest impact for your users so far?

qq. can devs adjust the threshold levels for features like background blur intensity or noise suppression aggressiveness via the API? by the way congrats👏 for launching @maxim_troshin

me enjoyed seeing so many effects in one SDK. Could developers enable only selected features to reduce app size? That would make integration more flexible.

How does this hold up performance-wise on mid-range mobile devices, especially when stacking multiple effects like background blur plus beautification at the same time?

How does the latency compare to cloud-based alternatives like Zoom's background blur, especially on mid-range mobile devices where processing power is more limited?

How well does it hold up on mid-range mobile devices, especially when stacking multiple effects like background blur plus noise suppression at the same time?

Does running all of this client-side mean there are noticeable CPU or battery drain issues on mobile, or have you found ways to keep performance smooth across devices?

How does the client-side performance hold up on lower-end mobile devices, especially when running background blur and noise suppression at the same time?

What's the typical latency for the real-time audio effects? Asking as someone who runs a live music stream.

Client-side processing is a huge win for privacy, and the API looks clean enough to integrate in an afternoon. The real-time noise suppression actually held up well on a noisy laptop fan during my quick test.

The client-side approach makes the integration story especially interesting because performance becomes part of the product experience, not just an infrastructure detail. Beyond the quality presets and per-effect controls, do you expose any runtime signal that helps an app decide when to reduce effect complexity or disable a feature on weaker devices? That kind of graceful degradation seems important for products with a wide range of hardware.

Congrats! Since everything runs client-side, curious what battery/CPU looks like on mobile once you stack a few effects together, like blur and noise suppression running at the same time.

How does the client-side processing hold up on lower-end mobile devices, especially for real-time effects like background blur and noise suppression together?

The fully client-side, nothing-hits-your-servers approach is what makes this appealing for a small team — no per-minute processing bill to worry about. As an indie dev building an audio-first app: can I pull in just the real-time noise suppression module on its own without the whole video pipeline, and roughly how much cold-start load does the WASM add on first use? Clean launch.

How does performance hold up on lower-end mobile devices when running something like background blur plus noise suppression at the same time, and is there any fallback when the GPU isn't capable?

Running everything client-side is a big plus. Privacy matters more than ever, so it's great to see features like background blur and noise suppression without sending data to external servers. I'd be interested to know how it performs on lower-end devices.

Since Effects SDK runs entirely client-side, how does performance hold up on lower-end devices when you stack multiple effects together, like background blur plus noise suppression at the same time?

Real-time is the word that makes or breaks this for me. I cut a lot of screen-recorded product demos, and the effects I care about (blur, denoise, auto-captions) are cheap to run after the fact but brutal live. What's the latency budget on the audio side before it drifts out of sync with the video? That's usually where "real-time" quietly becomes "near-time."

Client-side and on-device is exactly why I'd reach for this over a server-side pipeline — for a privacy-sensitive app it means the video frames never leave the browser. On the plumbing: what's the added bundle weight, and does the segmentation model require WebGPU or is there a WASM/CPU fallback for users on older machines? Trying to gauge how gracefully background blur degrades on a low-end laptop before wiring it into an existing WebRTC stack.

Amazing work, congrats on the release! I'm currently using your previous SDK for virtual backgrounds in my app AlterCam, and it's been a great experience. I'll definitely be checking out and testing the new Effects SDK soon. Best of luck on Product Hunt! 🙌

how long does it take it deploy this app?

Yeah Rohan! I consider myself as one of the Cloudflare's biggest fan and deploying just by dropping a folder is amazing. 60 minutes sound like too short but I'm sure it's more than enough. Thanks for hunting it

The dashboard layout is genuinely one of the cleanest I've used for managing DNS and security settings, everything feels like it's exactly where you'd expect it to be.

we run a handful of small static sites on cloudflare pages, all deploy on git push and the free tier hasn't blinked. Drop looks like an even lower on-ramp. the thing i'd actually want to know before using it for real: once i claim a Drop, does it become a normal Pages project i can wire git deploys + D1 into, or is it a separate one-off?

Been using Cloudflare's free tier for a side project and the speed boost was immediate, pages just feel snappier now.

Been using Cloudflare's free tier on a few personal sites and the DNS speed boost was noticeable right away. Also love how quickly I can flip on firewall rules without digging through settings.

Set up took maybe ten minutes and the dashboard immediately showed my traffic getting blocked from sketchy IPs I didnt even know existed.

How does the pricing actually scale for a site with moderate traffic - do you end up paying a lot once you cross into the next tier or is it pretty reasonable as you grow?

How does the pricing actually scale for a small site that suddenly gets a traffic spike, or do you cap usage before that kicks in?

The dashboard layout is genuinely one of the cleanest I've seen in the CDN space. Setting up DNS, firewall rules, and caching all from one sidebar without feeling cluttered took some real thought.

How does the free tier hold up under a sudden traffic spike, and at what point does it nudge you toward a paid plan?

Cloudflare has so many products bundled in now, where do you even start as someone just trying to speed up a small WordPress site and add basic DDoS protection without getting buried in enterprise features?

Finally tried Cloudflare for my small site and the setup was surprisingly painless. The dashboard makes it easy to toggle on the basic security and performance features without digging through docs.

Been using Cloudflare for a while now and the DNS speed bump alone was noticeable right away. Also love that the free tier covers way more than I expected for a small project.

Dropping a folder and having it live in seconds with no account needed is exactly the kind of friction that should be gone for quick projects or demos. My real question is the same one raised in the comments: if the site has a build step (package.json), does it actually run the build, or just serve the files as-is? That's what decides whether this is truly "instant" or only for simple static sites. The 60-minute lifespan also feels short, though I get it's there to prevent abuse.

how does the pricing actually scale for a small site that's starting to get more traffic?

The drag-a-folder-and-it's-live flow is exactly the friction I want gone for tiny side projects. What happens when there's a build step though, does it serve the files as-is or will it run a build if it sees a package.json? Most of my throwaway projects need one npm build before they're actually deployable, and that's usually where "instant" stops being instant.

This looks incredibly frictionless! With no account required and an instant live link, how is the team mitigating the risk of abuse, like rapid-fire phishing campaigns or malicious hosting? Are there automated security scans happening on the .zip files before the 60-minute URL goes live?

How does the pricing actually scale for smaller sites compared to big enterprise customers, and are there any hidden caps on traffic or requests on the free tier that might catch you off guard?

How does pricing actually scale for a mid-size SaaS doing around 50M requests a month, and are there any gotchas when integrating with custom WAF rules alongside your existing CDN setup?

Hi PH 👋

It's Theme Hospital meets RollerCoaster Tycoon, set in a lovingly legally-distinct San Francisco (Goggle, Chirper, Hacker Mews). The sim under the hood is real: raise prices and conversion drops, ship without QA and bugs breed, a rival will undercut you to "free, for now."

It's free, runs in the browser, no account. Every run has a seed you can share as a challenge. Would love your feedback! I'm here all day.

Happy Saturday!

I can already imagine founders saying "just one more run" and suddenly losing an hour. 😄

This is dangerously relevant for anyone building a startup :))

The dry announcer narrating every bad decision might be the best part. learning MRR, churn, CAC, burn rate, pricing, hiring, and shipping tradeoffs through a game sounds much more memorable than reading another SaaS metrics guide. "A rival will undercut you to free, for now" also feels painfully accurate.

I can already see myself hiring the dog before the company can afford it :)

As someone actually doing the startup thing, these games are either therapy or a personal attack, no in between. Does San Fran Sim model the boring failure modes (running out of runway, a cofounder walking, a launch that just flops quietly) or is it mostly the fun growth-curve fantasy? The unglamorous parts are the ones that would make me laugh.

played a quick round and the dry announcer calling out my churn mistakes had me laughing out loud. surprised how naturally it teaches MRR and CAC without feeling like a textbook.

Fun idea, I love these type of sim/manager games, but I think there could be more tool tips or help to explain what things mean

The shareable seed as a challenge is a lovely touch — Theme Hospital energy without any install friction. Since it runs in the browser with no account, does my run persist if I close the tab and come back to it later, or is the seed really the only way to pick things back up? Trying to work out whether this is a lunch-break-and-resume thing or best played in one sitting.

How long does a typical playthrough take, and does the dry announcer's commentary actually adapt based on the metrics you hit or is it kind of scripted throughout?

How realistic is the SaaS metrics simulation - does it actually factor in growth rate assumptions or just keep the numbers feeling random as things scale?

you can probably hook this up to the @ExploreYC API and get some real Y Combinator companies on the UI there - maybe an imaginative network hub with YC founders??

Played a round and actually learned something about CAC payback without realizing it. The dry announcer roasting my hiring decisions is a nice touch.

Ha, "the rollercoaster is your MRR chart" is painfully accurate. Gave it a quick go and it ran really smooth, nice implementation. I'm actually doing the startup thing right now, so it hits a bit too close to home. Gonna do a proper run in my free time. Congrats on the launch! 🚀

How deep does the simulation actually go on the metrics side - are churn and LTV calculated dynamically based on player decisions or do they follow fixed formulas?

I am having fun so far. Seriously cool vibe with the look and feel. Well done.

the dry announcer calling out your CAC miss while you're scrambling to ship a feature is a genuinely clever hook. love that it teaches SaaS metrics through actual pressure rather than a textbook.

Curious how realistic the SaaS metrics actually get here — like does churn scale with the decisions you make or is it more of a fixed curve so you can learn the formulas without it feeling random?

Interesting concept. I think that I have never seen some "startup" game done in the form of a roadmap. For me innovative! :)

The dry announcer narrating your missteps is such a clever touch, turns every failed launch into something you actually laugh at instead of rage-quit over. Love that it teaches MRR and churn by making you feel them first.

finally something that makes SaaS metrics click without putting me to sleep. the dry announcer roasting my runway decisions genuinely got a laugh out of me.

the dry announcer narrating your burn rate decisions is such a perfect touch, turns the usual dry metric lessons into something you actually want to keep playing. love that it runs in the browser with zero friction too.

How does the game handle late-stage scaling challenges, like managing a large engineering team or dealing with investor pressure once you approach IPO territory?

The dry narrator is genuinely funny and caught me off guard, especially when my churn spiked because I got cocky with pricing. Nice touch teaching MRR and CAC in a way that actually sticks instead of a boring textbook.

does the gameplay carry over between sessions or do you have to start from scratch in the garage every time you reload it

does the burn rate change based on which features you ship, or is it a flat meter the whole game

The dry announcer roasting my churn rate while I forget to refill the coffee machine is genuinely funny. Also didn't expect to learn what CAC actually feels like under pressure, school never did that.

Played a round and the dry announcer cracking jokes while my runway hit zero genuinely made me laugh out loud. The MRR and churn meters actually click once you watch them shift with each ship.

How realistic are the SaaS metrics in this - does the burn rate and churn actually follow plausible math, or is it more arcade-style?

The dry announcer narrating your dumb decisions is genuinely hilarious, and watching the MRR ticker climb while my burn rate quietly murders me is weirdly educational. Caught myself actually thinking about CAC vs LTV while playing, which is not what I expected from a browser game about a garage startup.

Really excited to hunt ChatGPT Work on Product Hunt! 🚀

A few things that stood out to me:

• 📱 Start a task on your phone and pick it up later on desktop or web.

• 💻 Desktop can work across local files, apps, and a built-in browser.

• 🔌 Connect tools like Slack, Google Drive, Gmail, Outlook, Teams, and Salesforce.

• 🌐 Turn prompts into interactive dashboards, trackers, and web apps with Sites.

• ⏰ Automate recurring work with Scheduled Tasks.

• 👨‍💻 Codex is now built into the desktop app for coding workflows.

Feels like a big step from "AI that answers questions" to "AI that actually helps move work forward."

Frankly not able to keep up with all new recent model updates and releases this week 👏

So we're officially from "AI assistant" to "AI coworker ". 😄 Now I just need it to survive Monday morning meetings too.

Finally got it to research a client and draft a proposal while I made coffee, came back to something I actually didn't hate editing. Curious how it handles messier workflows though.

the focus on letting the agent actually persist on multi-hour projects rather than just answering a single prompt feels like a really thoughtful design choice.

how does it actually keep working for hours without losing context or hitting some kind of usage cap

Finally something that actually follows through on a multi-step task without me babysitting it. Had it pull data from a spreadsheet and draft a summary doc end-to-end.

The agent actually stuck with a messy spreadsheet task for two hours and handed me a clean summary. Honestly didn't expect it to follow through that long without me babysitting it.

Had it pull together a client brief from scattered Slack threads and a Drive folder, and it actually handed back a usable draft in under an hour. The "stay with it" part feels real, not just marketing.

The way it sticks with a project for hours instead of bouncing you back every few minutes feels really well thought out. Most agents quit too early, so the persistence here is a standout choice.

Had it research competitors and draft a slide outline while I grabbed coffee, came back to something actually usable instead of a wall of placeholder text. The cross-app part feels genuinely new.

how does it handle permissions when it needs to act across multiple apps, do I have to approve every single action or can I set it to run longer chunks on its own?

Can it stick with recurring tasks (content scheduling, file cleanup) unattended, or does it need a human check-in each run?

The Salesforce connector is the line that caught my eye — WinBidIQ scores federal opportunities daily, and the obvious next step is pushing the ones worth pursuing straight into a pipeline tool instead of a spreadsheet. Building that matching logic myself taught me the write side is where it gets hairy: creating/updating the right record without duplicating or overwriting something a human already touched. Following on what Santino asked below about permission control — is there a review/checkpoint before ChatGPT Work commits a write to something like Salesforce, or does it just execute once it's decided on the action?

The shift from "assistant that answers" to "agent that carries a project for hours" is ambitious and powerful if it actually works. My biggest question, echoing others in the comments, is around permission control (what it can and can't touch) and whether the phone-to-desktop handoff truly syncs state or just restarts from scratch. Without a checkpoint to pause/review before irreversible steps, I'd be a bit nervous handing it real client work.

the fact that it can actually stick with a project for hours instead of dropping context mid-task is wild, feels like the first agent that respects long workflows

The phone-to-desktop handoff is the piece I'd test first: when I start a task on mobile and resume on desktop, does the in-progress state sync through my account, or does the desktop re-run it from the top? And do Scheduled Tasks fire server-side when my machine is asleep, or only while the desktop app is open? That local-vs-hosted line is what decides whether I can actually lean on the automation instead of babysitting it.

Nice one, been waiting for this. Very interested to see how token usage is.

How does it handle permissions when it needs to access files or apps across my workspace, and can I scope what it can touch so it doesnt go off and do something I didnt ask for?

The hours-long autonomy is exciting, but I'd love a way to step in and review or redirect the agent mid-task without killing the whole run. Something like a checkpoint that pauses for my input before a big irreversible step would make me way more comfortable handing off real client work.

Congratulations 🎊

I got tired of having to open/look at Visual Code to see how much I was earning on @Kickbacks.ai so I built a CLI and Mac Menu bar tool to easily see how much I was making, the ads I saw, and which days, weeks, and months I earned the most.

Feel free to leave feedback and hope folks enjoy :)

The compact two-line earnings view that lives in the terminal is a genuinely nice touch — one less dashboard tab to babysit. On the plumbing: does the CLI keep your Kickbacks token in the macOS Keychain or a plaintext dotfile, and does the menu bar app poll the API on an interval or hold a live connection for the real-time earnings?

love that you brought live earnings and ad history into the terminal and menu bar instead of forcing a dashboard login. the mac menu bar presence especially is a small touch that makes checking numbers feel frictionless.

Gabe, this is such a builder move :))

Getting tired of opening VS Code just to check earnings, then casually shipping a CLI and menu bar app for it... honestly very on-brand. I really like how focused it is too: status, live earnings, ad history, and simple trends without adding unnecessary complexity. The Mac menu bar version is probably what I'd end up checking way too often :)

Love that the menu bar app stays in sync with the CLI without a clunky refresh button, and the fact that I can pipe live earnings into other tools without copy-pasting from a dashboard. The little touches like a compact earnings view that fits in two lines of terminal real estate is a nice nod to people who actually live in their shell.

Finally a way to check my Kickbacks earnings without opening a browser. The menu bar widget is super handy and the CLI output is clean and fast.

finally a way to check my kickbacks without opening another browser tab, the menu bar widget is exactly what i wanted. nice work

Finally a way to check my kickbacks without opening yet another browser tab, the menu bar widget is genuinely handy when im working.

does the earnings data refresh in real time or pull on a set interval, and do you support any kind of alert when a payout is queued or approved?

how does it pull the live earnings data, is it querying the Kickbacks API directly or screen-scraping somewhere? curious if there is any rate limiting i would hit if i leave it running in the menu bar all day

How does the live earnings actually update, is it pulling from an API on a timer or streaming something real time? Also wondering what happens if you're offline for a while.

how does the menu bar piece stay updated without me needing to refresh anything, and does it use a lot of battery on macOS?

finally a way to check my kickbacks earnings without opening another browser tab, the menu bar widget is genuinely useful for someone like me who forgets to log in.

Love that this lives in the menu bar instead of yet another browser tab. The decision to keep it terminal-first for the actual data view feels thoughtful, especially for folks already living in a shell all day.

Does this pull data directly from the Kickbacks API or do I need to authenticate each time I open it, and is the menu bar version a separate download?

finally a way to check my kickbacks without opening another browser tab. the menu bar widget is genuinely handy, though i wish the cli command was a bit shorter to type.

how does the mac menu bar part sync with the CLI - do they pull from the same backend or do I need to configure them separately?

Does the menu bar version update in real time when new earnings come in, or do I have to manually refresh? Curious how chatty it would be running all day in the background.

the menu bar integration is such a clean touch, especially for creators who already live in the terminal all day

The 20-minute full-app trial you can restart forever is a good honesty signal, definitely better than the usual crippled-feature demo. Because it lets people prove it works on their exact setup before paying. Congrats on the launch!

The visible-wire routing with live meters on every channel is what would pull me off BlackHole — hand-wiring multi-output devices in Audio MIDI Setup, you can never actually see what's flowing. I'd use this mostly to record calls and screen captures: do the routes I set up persist across reboots and show back up as a stable input in OBS or Zoom on next launch, or is it a rewire each session?

I love the visual wire concept reminds me of Reason or old-school analog patches, qq does the app support multi-channel audio output (like 5.1 or 7.1 mapping) ? congrats to ship @chrisbattarbee

The visual routing approach makes this much easier to understand than traditional audio setup.

Does this work with Discord and OBS at the same time, or do I have to pick one app to receive the routed audio?

Finally a clean way to route my mic into OBS without the messy workarounds. Set it up in under a minute and it just works.

Finally something that routes my mic into OBS without a mess of cables or paid plugins. Setup took about two minutes and my stream audio just works now.

How does this handle latency when routing real-time mic input through to other apps, especially if I'm running multiple audio streams at once?

Finally a clean way to route mic input into apps that don't normally accept it, took about two minutes to set up and just worked. Super handy for screen recordings.

finally something that lets me route mic input into apps like Logic without needing a mess of virtual cables. setup took like two minutes and it just worked.

The way it just shows up as a proper audio device in macOS feels really well thought out, no clunky workarounds or hidden menus to dig through.

How does this handle latency, especially when routing mic input to other apps in real time?

The clean way it shows up as a virtual device in Audio MIDI Setup feels really considered, like the team actually thought through what Mac users expect. Nice execution on something that could have been clunky.

Finally a simple way to route my mic into OBS without the audio routing headache I usually deal with on Mac. Setup took about a minute and the latency seems low enough for calls.

Curious how the audio actually routes between apps, is it something like an aggregate device or do I pipe the mic through its own channel. Trying to picture the setup before installing.

How does this play with apps that already have audio routing built in, like Zoom or OBS?

How does this handle latency in real time, like when you are routing a mic for a live stream or podcast recording?

The menu bar icon is a really nice touch, super minimal and fits right in with macOS. Love how lightweight it feels compared to other audio routing tools I've tried.

As a producer this solves a real headache — getting a standalone synth or an iPad music app straight into Ableton without a hardware loopback cable. Quick q: can I route a DAW's master back into an input channel on the same machine for internal resampling, or does feeding an app into itself get weird? The sub-15ms latency + per-channel meters look perfect for that.

Been wanting proper per-app audio control on the Mac for years, the one system slider for everything drives me up a wall. Does SoundPipe let me route a single app's output on its own, like sending Kontakt to one interface while everything else stays on my headphones? I run virtual instruments and the routing is the part macOS makes weirdly hard.

Curious how this handles latency when routing mic input live into something like a Zoom call? That kind of sync issue would be a dealbreaker for me.

Coming from a music background, I'd love to pipe clean app audio (a backing track or score playback) straight into a video call for remote lessons, instead of it going through the mic. Is that a solid use case for SoundPipe? The one-click driver with no kext is a big plus.

How well does the AI handle complex joins across multiple tables, or does it get confused once you move past simple single-table queries?

Finally played around with Basedash after hearing about it for weeks and the natural language chart builder honestly surprised me the most. Plugged in a Postgres db, asked for weekly churn, and got a clean line chart in seconds.

How does it handle complex joins across multiple tables when you describe them in natural language, and are there limits on how many rows it can process per query?

Connected my Postgres database in under a minute and asked it to show churn by signup month. Got a clean chart back with a surprisingly useful written summary. Wish more BI tools felt this straightforward.

Connected our Postgres and asked it to show weekly active users broken down by plan. Got a clean chart in seconds and could refine it with a follow-up sentence. Way easier than firing up Metabase for a one-off question.

How does it handle messy or inconsistent schemas when you point it at a real production database, or do you need to clean things up before connecting?

How does the AI handle data sources it has never seen before, like a weird internal database or a spreadsheet with merged cells and weird formatting? Curious how forgiving it is in practice.

How does the AI handle messy or poorly documented schemas when you point it at a database it has never seen before?

I really like how Basedash turns a plain sentence into a working chart, that no-SQL magic trick feels impressively smooth and actually useful.

how does the pricing scale as the dataset grows, and are there any limits on the number of rows or sources you can connect before the AI starts getting sluggish on the queries?

connected my postgres db and asked it to chart weekly active users and the chart actually matched what i'd write by hand. the natural language piece feels less gimmicky than i expected.

the SCIM + RBAC/RLS separation is a clean call. a lot of tools conflate identity provisioning with access control and it creates a mess when someone departs mid-project. curious how you handle partial deprovisioning for contractors who are tied to active dashboards, any edge cases there? congrats on #6, 24 launches in is impressive staying power.

The place access always rots for us is offboarding, not onboarding. A contractor leaves and three tools still have them active two months later. Does the SCIM here cover clean deprovisioning across downstream apps, or is it mostly about pushing new users out? Curious how it handles the half-connected tools that don't speak SCIM properly.

The bit that decides a rollout for me is deprovisioning latency: when someone leaves or switches teams in the directory, does Basedash cut their access on the next SCIM sync cycle, or is there a near-real-time path? For a team standing up dashboards on live customer data, the gap between 'left the company' and 'can still open the dashboard' is the whole risk. Also curious whether group-to-role mapping is 1:1 or if you can express nested/combined groups.

How does the AI handle more complex joins across multiple data sources, and is that included in the base plan or does it require a higher tier?

Connected my postgres database and asked for a weekly active users chart - got a clean visualization in like 30 seconds without touching any SQL. The natural language piece actually works better than I expected.

Pulled it up to poke around and the natural language chart builder actually understood my messy request on the first try, which surprised me. Setup to connected data was quick and the dashboard looked clean without any fiddling.

Interesting to see a natural language interface paired with enterprise provisioning. It makes me wonder how the learning curve differs for new team members versus seasoned analysts when they first use it.

Connected my postgres db and asked for a chart of weekly active users, got a clean visualization in seconds. The natural language piece feels polished, not just a wrapper.

How does it handle complex joins across multiple tables when you ask in natural language, does it just guess at the relationships or do you need to define them upfront?

One thing I've noticed with enterprise features like SCIM is that the technical implementation is often the easy part compared to getting companies to prioritize identity management before it becomes painful.

Have you found SCIM to be something prospects actively ask for, or is it mostly a capability that helps remove blockers later in the sales process?

Congrats on the launch!

Hooked up our Postgres and asked it to chart weekly active users by plan, got a clean visualization back in seconds without touching SQL. Impressed it actually understood the relationships between tables without me hand-holding it.

Would love to see a way to save and share prompt templates for common chart types. Our team keeps asking similar questions each week, and being able to save a "weekly revenue by plan" prompt with preset filters would save a lot of time.

@maxmusing @kris_lachance The offboarding path is the part most teams underestimate, that's where access quietly rots. Auto-syncing group and org membership on every change, not just provisioning new users, is the hard part done right. That's exactly what makes "who still has access" answerable. Nice work.

the split between org membership and resource access via RBAC/RLS is the right seam. most SCIM implementations conflate the two and enterprises pay for it later.

real q: does the audit log preserve who was in the org at the moment a query ran or a dashboard was published? asking because compliance teams increasingly want "this insight was produced by this human on this date under these roles" as a first-class record and most tools reconstruct it after the fact.

The natural language chart builder actually worked on the first try with a messy Postgres schema, which caught me off guard. Wish the dashboard sharing was a bit more polished though.

Hi Product Hunt! 👋

I'm Artur, maker of Breathing in Labour.

This started when my partner and I were preparing for birth: breathing is the

one tool every birth class teaches, but during a contraction nobody remembers

"inhale 4, exhale 8." You need something to follow, not something to remember.

So I built a breath coach you follow, not read — a calm animation paces every

inhale, hold and exhale. It was born for labor, but the same techniques work

for anyone: stress, anxious moments, falling asleep, box breathing at 3am with

a newborn. If you breathe, it's for you. 🫁

The part I'm most proud of: everything you actually need is free. All 7 guided

techniques — Box, Up, Relaxation, "Haa" and Candle Blowing for early labor and

everyday calm; Down and J-Shaped for the pushing stage — free, forever, no

exercise paywall.

📖 It teaches, not just paces. Every technique comes with the full context:

what it's for, when in labor to use it (early Stage I? between contractions?

pushing?), step-by-step instructions and a visualization to hold onto — plus a

guide on why breathing matters and how to practice, including together with

your birth partner. You walk in knowing why you're breathing this way, not

just following shapes.

Most breathing apps lock the breathing itself behind a subscription. I did the

opposite: Premium only changes how the app looks and feels — audio soundscapes,

color themes, and building your own custom breathing patterns. The core

practice is never the product being sold.

🎁 Want to try Premium anyway? We made a launch offer just for Product Hunt:

open lebene.pl/en/producthunt and you get the first 3 days of the subscription

free (on iPhone the code redeems automatically in the App Store; on Android

install first, then open that link).

One honest tip: the offer starts a real subscription with the first 3 days

free. If you just want to test it, cancel the subscription in your store

settings right away, then go explore — you keep Premium for the full 3 days,

and there's nothing to forget later. Valid through July 18.

I'd love feedback from everyone — parents, doulas, midwives, and anyone who's

ever needed to catch their breath. What would make this more useful for you? 💙

finally an app that skips the ads and accounts nonsense, just straight to the breathing. the visual rhythm guide actually helped me slow my breath down quickly.

finally something that actually feels calming instead of overwhelming, the minimalist rhythm visual is genuinely soothing during practice

Does the app actually adapt the breathing pace to whatever stage of labor someone is in, or is it more of a static guide you have to manually switch between?

Love the no-ads approach, super refreshing for a birth prep app. Does the app guide you through each stage automatically or do you have to manually switch between early, active, and pushing phases during a contraction?

How does the app actually guide you through the different stages, like is there audio cues or just the visual feedback to follow along with?

The minimalist interface really does feel calming, and I love that the core breathing techniques are free without any sign-up hoops. The visual rhythm feedback helped me actually focus on slowing my breath instead of getting distracted.

How does it actually guide you through the breathing in real time, like is there audio cues or just the visual feedback to follow along?

The breathing visuals are really calming, especially during the contraction timing exercises. Wish more apps followed this no-ads, no-account approach for something this personal.

The minimalist interface actually made it easier to focus on the breathing rhythm, which is what you need mid-contraction. Nice that the core stuff is free without forcing a sign-up.

the choice to skip accounts and ads for something as vulnerable as labor prep shows real care for the user. love that the visual breathing feedback is the whole interface instead of buried under menus.

The minimalist design really helps you focus on the breathing rhythms without distractions. Surprised how calming the visual feedback feels during practice.

The minimalist visual feedback for breathing rhythms is such a thoughtful touch. Calm, uncluttered, and actually useful when you're mid-contraction and need zero friction.

Does the app guide you through each contraction in real time during labor, or is it more of a practice tool you use beforehand to build the breathing habits?

No ads, no forced account, free core mechanics on a labor app is a real flex. That whole category usually holds your breathing hostage behind a signup, so good on you for not doing it. My actual question is about the moment of use. Mid-contraction, nobody is studying a screen. How does the visual feedback carry when someone can barely look at it? Curious whether you leaned on haptics or a sound cue so the rhythm still lands with eyes closed.

First off Congratulations on launching the app! When I used the it the UI felt like it is genuinely well thought and does calm people down. The mock up labor stages are a great way to prep yourself for the real deal. I have used many meditation apps before and I have to say this was the most simple yet effective.

The minimalist interface really helped me focus on the breathing patterns without distractions, and the calming visual feedback makes practicing feel almost meditative. Nice to see a labor prep app that respects your time and attention.

The minimalist interface really helped me focus on the breathing rhythms without distractions. Loved that the core features are free with no signup required, makes it easy to actually practice daily.

How does the app actually guide you through the different stages of labor in real time, since labor itself is pretty unpredictable in length and intensity?

How does the app actually cue the breathing patterns during a contraction when you might not be able to look at the screen, especially if the visual feedback is the main guide?

The minimalist interface actually made it easier to focus on my breath instead of fiddling with settings. Surprised the core features are genuinely free with no sign-up walls to jump through.

Hey, I'm Francesco, one of three people behind Trailsense, along with Felix and Lea.

We build small solar-powered sensor nodes that count how many people pass by on a hiking trail, without cameras and without hikers having to install anything. The nodes listen for wifi probe requests, the background pings a phone sends out while scanning for known networks, and turn them into an aggregated count. No raw probe request data ever leaves the node, and no personal data is stored anywhere in the system.

The idea started at the Tourism Technology Hackathon in 2025, where Tirol Werbung brought us a real problem: they know exactly how many people sleep in which hotel each night, but almost nothing about where those people go during the day. We kept working on it as our master's project, and this year we ran a pilot on the Nordkette in Innsbruck alongside Tirol Werbung.

Right now Trailsense is pre-commercial. The pilot is done, we're refining the dashboard and the node hardware, and we don't have a public demo up yet since it's still a bit rough around the edges. Longer term we want to add a public-facing side too, something like Google Maps traffic data but for hiking trails, so hikers can check how busy a trail is before heading out. For now our focus is the municipality and tourism board side.

We're posting here mainly to get feedback, whether you've built sensor hardware yourself or you're just someone who hikes and has opinions about being tracked outdoors. What would make you trust a system like this to collect data around you? And if you were a hiker using a public facing traffic-map version someday, what would you actually want to see on it?

Our node firmware is open source on GitHub (github.com/trailsense) if you want to see exactly how the data collection works.

Happy to answer any questions.

Finally launching TrailSense today!

I really enjoyed working on a product that includes hardware. It was my first time being involved in something like that, and it made the whole project especially exciting.

Also here for any questions about our product :)

Took a look at the dashboard and the wifi probe approach is clever - feels like the trailhead counts line up with what I'd expect on a busy weekend. Nice that nothing leaves the device.

The no-app approach is clever and the wifi probe trick actually makes sense after reading about it. The dashboard view over time looks like the kind of thing I'd check before a weekend hike.

clever way to get a busyness signal without cameras or an app install. one thing i'm curious about given how the tech works - modern phones (iOS and most recent android) randomize their MAC address per probe request specifically to prevent exactly this kind of passive counting. does trailsense do any signal-strength/timing clustering to guess that five different randomized addresses over an hour are probably the same phone, or is the count more of a relative busyness proxy (more pings = busier trail) rather than an actual hiker headcount? asking because the accuracy story is pretty different depending on which one it is.

finally a privacy first take on trail counters, no app or camera nonsense is a big plus. curious how reliable the wifi probe counts actually are in dense spots.

How does it actually pull the numbers in? Do I have to type everything out each month or can it connect to Stripe or a spreadsheet and draft the update for me?

Super clean setup, got my first update drafted in under 10 minutes. Love that the templates nudge you to share failures, not just wins.

Building the list once and reusing it is the real friction-killer here — the copy-paste-to-20-people step is exactly why updates slip. One thing I would want before making this my monthly ritual: can I segment a single send so a specific metric or paragraph only reaches a subset (lead investors vs smaller angels), or does every recipient get the identical email? That is what decides whether I run one send a month or quietly maintain two versions by hand.

The template saved me from staring at a blank doc for an hour, and pulling in metrics from our dashboard actually worked on the first try. Wish I had this during my last round.

Honest take after a quick try: the templates actually sound like a founder wrote them, not a robot. Wish I'd had this when I was juggling 12 angels and a Google Doc.

bulk-send with per-recipient tracking is the right shape for this. curious about one thing not covered yet: do you support sending a different level of detail to different investors, like your lead gets the full numbers and a smaller angel gets the summary version, or is it genuinely one update broadcast to the whole list every time? that's usually where founders end up back in manual copy-paste mode even with a tool like this.

How does this handle the actual writing part though, or is it just a template and distribution tool I have to draft the content myself each month?

Does this pull metrics automatically from tools like Stripe or QuickBooks, or do I have to type everything out by hand each month?

I like that this is centered around investor communication rather than fundraising itself.

One thing I've always wondered is whether the biggest challenge is actually writing the update, or deciding how transparent to be when growth stalls or milestones slip.

Has that been a pattern you've noticed while building this?

Congrats on the launch! 🚀

Love how clean and focused the whole flow is. The fact that it nudges you to actually cover failures and hires alongside the wins feels like a real founder-friendly touch, not just a fancy mail merge.

A monthly update template that pulls metrics automatically from tools like Stripe or Mixpanel would save founders a ton of time. Most of us spend an hour formatting numbers each cycle when that data should just flow in.

The template structure made it so much faster to send our last update, and my lead investor actually replied saying it was the clearest one yet.

Finally tried a few templates and the monthly format actually pushed me to track metrics I had been ignoring. The "failures" section felt awkward at first but ended up being useful for keeping investors in the loop.

Curious how this differs from running something like Lighthouse or Ahrefs site audit, since those already flag a lot of trust and SEO issues. Is Plexa more about consolidating the noise into a single prioritized list, or does it actually catch things those tools miss?

How does Plexa handle sites behind authentication or ones with heavy client-side rendering, and does the scan run on demand or on a recurring schedule?

Finally tried Plexa on our company site and was honestly surprised how plainly it explained the SSL chain issue we had been ignoring for months. The prioritized list of fixes made it easy to know what to tackle first without a developer.

respect the rebrand story, shipping past the honest feedback instead of ignoring it is the harder path but the right one. question on the trust badge specifically: if a site fixes an issue right after a bad scan, does the public verification page update in near real time, or is there a lag where a visitor could see a stale "verified" badge on a site that currently has a live issue? that gap is the thing that would worry me most about any public trust score.

Building Plexa has been an incredible journey, and today’s feedback has already given me so many ideas.

To say thank you, I’m giving everyone from Product Hunt 1 month of Plexa Pro free.

If you decide to give it a try, I’d love to hear what you like, what you don’t, and what you’d build next.

Use code PRODUCTHUNT during checkout.

Thanks again for making launch day so much fun!

the way plexa frames every issue with the why behind it feels really considered. most tools just dump a list of errors and leave you guessing, so this actually stands out.

Finally tried Plexa on my portfolio site and it caught three SSL issues I had no idea about. The plain-language explanations actually make sense without needing a dev background.

A scheduled scan option would be killer, so I can set it to run weekly and just get a digest email of any new trust or SEO regressions instead of remembering to check manually.

When you say white-label is included, does that mean clients would see our branding throughout, including in the emails and CRM they use, or is there still a Market Genie footer somewhere?

The email writer saved me a solid 20 minutes on my first campaign and the tone actually matched my brand voice surprisingly well. Wish the funnel templates were a bit more flexible, but overall a strong update.

Does the lead scraper pull from platforms like LinkedIn and Instagram too, or just specific directories?

Congrats on your launch, Darrell.

As someone who has used Market Genie, I can only recommend it. It has everything you need in one place without burning your wallet. And it's so smooth to use.

Amazing work!

How does the white-label side actually work in practice, can my clients log into something branded as my own agency or is it more of a reseller setup?

How does the white-label actually work in practice, can my clients log into a fully branded portal or is it more of a resold account?

The visual automation builder is way cleaner than I expected for an all-in-one tool, and the lead scraper actually pulls useful contacts without a ton of setup. White-label at zero cost is wild for founders.

Spent an afternoon poking around the funnel builder and was surprised how smooth the visual automation felt, no clunky drag-and-drop delays like I expected. Curious to see how the lead scraper holds up at scale.

the local-only storage with optional AI processing is a really thoughtful split, it means you actually own your screen history without locking yourself out of useful analysis. nice to see a tool that respects privacy without gutting the feature set.

tried it for an afternoon and the timeline view feels genuinely useful, scrolling back to find that one design file i lost track of. like that the recordings stay on my machine by default.

How does the local storage hold up over time, like does it balloon fast on a normal workday and need manual cleanup or does it handle that itself?

Comes in handy when I can't remember which file I had open two hours ago, just ask it in plain English and it pulls up the right window. Love that the recordings stay local by default.

The timeline view is genuinely useful for finding that one thing I worked on last Tuesday without digging through screenshots. Local-only storage is a nice touch too, makes it feel less invasive than other activity trackers I have tried.

The "AI analysis is optional and uses your own provider or a compatible local endpoint" line is what makes this usable for privacy-sensitive work — the timeline stays local and I decide if anything ever leaves. Two things I would test before leaving it always-on: can I exclude specific apps or windows (a password manager, banking) from being recorded at all, and does the searchable text index live in the same local store, or does querying round-trip to the provider? Trying to gauge whether I can keep it capturing during sensitive work without leaking window contents into an external model.

Tried it for a day and the natural language search worked way better than expected. I asked "show me when I was working on the pricing page" and it pulled up the right window. Local storage is a nice touch.

How does the search actually work under the hood, does it index OCR text locally or rely on screenshots only when you query?

The timeline view actually made my scattered work day feel coherent, and being able to ask "what was that thing I saw yesterday" in plain English saved me a real hunt through screenshots.

This looks genuinely useful. I'm constantly trying to find a tab, file, or random thing i saw earlier and can't remember where it was.

Also really nice that everything stays local. Doessearch still feel fast once you've built up a few months of history?

the "recordings stay local, AI analysis is optional and uses your own provider" split is the interesting design decision here. if someone plugs in a remote provider (say a hosted API instead of a local endpoint) to ask a plain-english question, does the relevant screen evidence get sent up to that provider to answer it, or does it stay local and only the extracted text/metadata leaves the machine? for a tool whose whole pitch is private visual memory of everything you do, that's the detail that actually determines whether it's private by default or private-if-you-configure-it-carefully.

Finally something that fixed my clipboard mess without making me sign up for yet another account. The auto-grouping actually works well when I'm juggling snippets between Slack and my editor.

how does the secret masking with encryption actually work in practice?

Finally something that solves my clipboard chaos on Linux without making me log in. The auto-grouping is surprisingly smart, it actually separated a recipe from a hex color I copied moments apart.

Totally there with you on the need for better clipboards, one serious want of mine is to have cross device (even cross OS) clipboards. Have you looked into a solution for that at all?

finally a clipboard manager that actually feels private. the local-only setup and secret masking sealed it for me, and full-text search across old snippets is a small thing that changed how i work daily.

Curious how the search holds up when the paste stack gets into the thousands — does it stay snappy or is there a noticeable slowdown over time?

tauri + sqlite fts5 + no electron is a good sign for something that's meant to sit quietly in the background all day. one question on the secret masking - since it's presumably detecting secrets by pattern (known token/key shapes) rather than a hardcoded list, what happens to an arbitrary sensitive string that doesn't match a recognized pattern, like a random internal password someone pastes that isn't formatted like an API key? does it fall back to encrypting everything at rest regardless of whether it was flagged as a secret, or is the AES-256 specifically reserved for things that got detected? the headline feature is "secret masking" so i'd want to know if there's a safety net under it.

finally a clipboard manager that doesn't try to sync to some random cloud. The full-text search over old copies is genuinely useful, found a snippet I'd lost weeks ago in seconds.

Been hunting for a clipboard manager that actually keeps images and files intact, and this one nailed it. The local-only encryption for sensitive stuff is a nice touch I didn't know I needed.

Finally tried it last night and the auto-grouping actually caught me off guard — it just figured out what belonged together without me tagging anything. Local-only and no account is the cherry on top.

The "fitting room on every store, works on any site" angle is way more useful to me than the single-retailer try-ons that only work where they are already built in. Day-one question: do I set my photo up once and it is reused across every site, or is it a fresh upload per garment — and can I save or share a render, or does it vanish when I leave the page? Trying to work out whether this fits a real "send the look to a friend before I buy" moment.

Finally tried this on a jacket I was eyeing and it actually looked realistic, not like a rough sticker. The ten second claim feels accurate too.

does it work on every site or do stores have to opt in first

"processed in memory, never stored" is a good privacy line but it leaves one thing ambiguous: does that memory processing happen locally in the browser, or does the photo get sent to a server to generate the photoreal render and just not get persisted afterward? those are very different privacy claims even if the end result (nothing stored) is the same, and "never leaves your device" earlier in the pitch reads like the former.

Tried it on a couple of sites and it actually delivered, the preview popped up while I was browsing Nordstrom and the fit looked right. Way more useful than I expected for a quick checkout decision.

Tried it on a few clothing sites and it actually worked faster than I expected, like under 10 seconds for real. The try-on preview looked pretty natural on me, not uncanny at all.

Tried it on a couple of sites and it really did pop on me in like 15 seconds, which honestly surprised me. The fit looked pretty natural too, not stretched or warped like other try-on tools I have tested.

Finally a try-on tool that actually works on random sites, not just the brand's own page. The fit on a linen shirt looked close enough to my shoulders that I'd trust it for a first-pass yes or no.

Congrats on launching! The 10-second claim is bold, how consistent is the output when the product photo is low quality or shot at a weird angle? Fashion sites vary wildly in image quality, so I'm curious how it handles the messy real world. Also smart choice making it a Chrome extension instead of yet another app; meeting people where they already shop makes a lot of sense.

Tried on multiple websites it removes the hassle of returns. Nice product saved a lot of time

Being able to see how a piece looks on me before buying would honestly save me from so many returns. One thing I'd love is a side by side comparison mode where I can try two or three garments at once on the same page, makes it way easier to actually decide which one to go with.

Tried it on a few shirts on a random site and it actually worked faster than I expected, the fit preview was close enough to my usual size to be useful.

Hi Product Hunt! 👋

A few weeks ago I left Claude Code running overnight. In the morning: 2M tokens gone, and the "fix" was editing the same file in a loop for 40 minutes. I had no idea *why* — the evidence had scrolled away.

Then I realized: Claude Code already writes a complete transcript of every session to disk. Nobody was reading it.

agentfdr is the crash investigator's toolkit for those recordings:

🛫 A per-turn timeline of every tool call, the context window's composition, and output tokens
🚨 Automatic anomaly flags: tool loops, error streaks, context bloat, cache thrash
🔍 A dissection panel for any turn — what it saw, what it did, what it cost
📊 Plan usage (5-hour window / weekly) with calibratable budgets
🚦 A CI gate: `agentfdr assert --no-loops --max-tokens 2M`

Zero instrumentation — `npx agentfdr` and you're looking at your own sessions. Nothing leaves your machine. MIT licensed.

Zooming out: working with coding agents is quickly becoming loop engineering — you design and operate the loop your harness runs: what enters the context, which tools fire, when it stops. You can't engineer a loop you can't see. agentfdr wants to be the instrument panel for that discipline.

Fun fact: every screenshot in the gallery is agentfdr dissecting *the session in which Claude Code built agentfdr itself* — flags and all. 🛬

I'd love your feedback — especially which agent CLIs to support next (Codex CLI? Gemini CLI? Aider?).

finally something that surfaces loop patterns in long claude sessions without making me grep through jsonl files, the live watch mode is genuinely useful when a run starts drifting.

How does this handle Claude Code sessions that span multiple branches or get interrupted mid-turn? Wondering if the timeline stays coherent or if the anomaly detection gets noisy when there's a context reset.

how does it flag loops in real time without hooking into the actual conversation state

finally something that makes my long claude code sessions actually reviewable, the loop detection caught a stuck tool call i would have missed for sure

does the anomaly detection work offline too, or does it need to ping anywhere to flag loops and token burns? curious how it stays 100% local with that kind of analysis.

Love that you skipped the SDK and just read what Claude Code already writes to disk, zero instrumentation feels like the right call. The per-turn timeline plus automatic loop and token burn flags is exactly what I'd want when something goes sideways mid-session.

How does it actually surface the anomaly flags in real time during a session, and is there any way to set custom thresholds for things like token burn before it pings you?

the "assert --no-loops --max-tokens" gate is the part i'd actually want running unattended, not just the timeline after the fact. question on tuning it though: some loops are legitimate, retrying a failing test until it passes, polling a build, re-reading a file after an edit. how does it tell that apart from a genuinely pathological loop before it kills the run? is the anomaly detection just repetition-count based, or does it look at whether the tool calls are converging toward something (diffs shrinking, errors changing) vs just spinning in place

how does it pick up the transcripts — is there a config to point it at a custom claude code logs directory or does it default to ~/.claude?

pulled comments from a few of my own videos and the CSV export was clean enough to drop straight into a spreadsheet, which saved me the usual copy-paste headache.

Love how the export formats and filtering options are all laid out without burying them in menus. Most scraping tools make you dig through settings just to get the basics, this one just works right out of the gate.

The no-code filter and search combo is genuinely useful, most exporters just dump everything and leave the hard part to you. Love that it handles multiple formats out of the box too.

Would love to see a sentiment analysis column added to the export, so I can quickly sort comments by positive, negative, or neutral without having to run them through another tool afterwards.

How does it handle videos with thousands of comments, does it cap the export or pull everything in one go? Also wondering if the filtering works before or after the data is fetched.

How does it handle videos with thousands of comments — is there a hard cap per video, or can it pull the full thread including replies?

The multi-format export option is a really thoughtful touch, especially letting people skip the API setup entirely. Going to use this for pulling audience feedback on client videos.

Pulled comments from a few videos and the filter by date saved me a ton of scrolling, way faster than coping with the default YouTube UI.

skipping the API-vs-scraping question since Dennis already asked it and that's the one I'd want answered first too. separate question: does the export preserve reply threading (which comments are replies to which), or does it flatten everything into one list? that matters a lot for the "product ideas and customer questions" use case, a reply only makes sense in context of what it's replying to, and flattened it just reads as a random opinion.

Congrats on the launch! Genuinely curious about the technical side: are you pulling comments through the official YouTube Data API or scraping them? I'm asking because the API has pretty strict daily quotas, and scraping tends to break whenever YouTube changes their frontend. How are you handling rate limits for videos with tens of thousands of comments? Would love to hear how you've thought about the reliability side, that seems like the make-or-break factor for a tool like this.

How does the scoring actually work when you pull in competitor data, are you comparing against the same eight categories or a different benchmark set?

Ran it on my small site and the breakdown actually highlighted a few gaps I had missed, especially around AI discoverability. Took maybe three minutes and felt genuinely useful rather than generic.

How does this handle businesses that operate in multiple locations or industries at the same time, or is it strictly single niche?

Took the free assessment and was surprised how specific the recommendations were, especially around my Google Business Profile gaps. The score breakdown made it easy to see where to focus first.

How does the AI discoverability check actually work, since AI tools vary so widely in how they surface information?

how does the visibility score actually stay current, like does it rerun automatically when my site or reviews change or is it a one-time snapshot?

Curious how the score actually weights the different pieces—does AI discoverability count as much as the Google Business Profile stuff, or is it more of a balanced blend across all eight areas?

How does the AI discoverability scoring actually work, does it check how my business shows up in ChatGPT and Gemini responses or something different?

Ran the free score on my site and was honestly surprised how it broke down AI discoverability alongside the usual SEO checks. The recommendations felt specific rather than generic, which is rare for these tools.

love how the visibility score breaks down into clear categories instead of one vague number, makes it actually actionable.

Have you considered adding a competitor comparison view? Something where you can see how your Visibility Score stacks up against two or three nearby businesses in the same industry would be really helpful for understanding where you actually stand and what to prioritise fixing first.

Would love to see a competitor benchmarking option so the visibility score shows how you stack up against similar local businesses, not just your own metrics in isolation. That context would make the recommendations way more actionable.

How does the AI discoverability piece actually work in the scoring — are you checking how brands show up in ChatGPT and Gemini responses, or something else?

How does this compare to something like SEMrush or Ahrefs for the SEO and visibility parts, especially if someone is already paying for those tools?

Love that the scoring covers AI discoverability alongside the usual SEO and review checks. Most visibility tools still treat AI search as an afterthought, so actually weighting it in the assessment feels like a smart move.

How does the AI discoverability scoring actually work under the hood, and does it need my API key for every single assessment or just the first one?

the machine-crawlability framing on the AI discoverability check makes sense as a proxy, but crawlability and actually getting cited in an AI answer are two different things - a perfectly crawlable site can still lose to a competitor the model just happens to favor for reasons that have nothing to do with schema or structure. do you have any way of validating the AI discoverability score against real-world outcomes, like actually querying ChatGPT/Perplexity with relevant prompts and checking if the business shows up, or is it purely structural signals for now?

Took the assessment on a whim and was impressed by how it broke down AI discoverability alongside the usual SEO stuff, which I haven't seen other tools cover well. The recommendations felt actually doable.

How does the on-device text recognition handle handwritten notes or low-quality photos with shadows, or is it strictly for clean printed pages?

How does the on-device text recognition hold up with longer or handwritten documents, and is there any plan to support more languages beyond English, Russian, and Romanian down the line?

the on-device search actually works shockingly well, found an old receipt in seconds. love that there's nothing to sign up for or unlock.

Downloaded this on a whim and genuinely surprised by how fast the edge detection works, even on crumpled receipts. The on-device text search is what sold me though, finding an old contract in seconds without it leaving my phone.

Scanned a few receipts and the edge detection picked everything up perfectly without me adjusting anything. The on-device search actually finds text in old Russian documents I had, which is a first for me.

Love the no-subscription stance, that's a breath of fresh air for scanner apps. One thing that would make this a daily driver for me: support for splitting a multi-page scan into separate documents automatically, or at least let me reorder and delete pages right inside the app before exporting the PDF.

the "everything stays on your phone" pitch is great until you factor in what happens when the phone dies or gets replaced. no cloud is a real privacy win but it also means no cloud backup by default, is there any local export/backup path (iCloud folder, encrypted file, anything) so a lost phone doesn't mean losing every scanned document, or is that an intentional tradeoff you're fine with?

love that this stays fully on-device — on-device text search across multiple languages without an account is genuinely rare, and the no-watermark pdf export is a quiet win for anyone who actually uses these daily.

How do you handle the real-time pricing for less liquid GCC stocks or gold from local dealers that don't have a public feed? Curious if I'm stuck entering prices manually.

mine's vested startup equity, options with a 409A that's basically a made-up number until there's an actual exit or a secondary sale. every net worth tool either ignores it completely or lets you enter it as a static value that's stale the day you type it in, since there's no live price to pull. curious if that's even something you'd want to touch, since "cost basis and honest P&L" gets a lot harder when the asset in question doesn't have an honest current price to begin with. the no-broker-login, log-it-yourself approach at least seems like the right foundation for it though, unlike anything that expects to sync from a live market feed