PH热榜 | 2026-08-22

@dmitriychuta Tracking App Store rankings across 175 countries with no API keys or App Store Connect access is the part I'm most curious about the App Store doesn't expose an official public rankings API as far as I know. Is this built on the public storefront pages themselves, and if so, how do you keep 175 countries' worth of that in sync 24/7 without it becoming a maintenance nightmare every time Apple tweaks something?

@dmitriychuta Love that it works with zero API keys or setup!

How are you getting reliable data across 175 countries without official access?

@dmitriychuta This solves something we've been doing manually for Cove — tracking how ranking moves after each App Store listing change (we've been testing subtitle variants and reordering screenshots). The no-API-keys, no-App-Store-Connect-access setup is the part that'll actually get me to use it instead of another spreadsheet. Adding Cove now to see how it holds up.

@yelyzaveta_kibets Thanks for the great question! The app won’t bombard you with notifications.

We track ranking changes for every app, and you’re right that apps ranking in many countries can generate a lot of activity. We call these changes “Events.” All Events are processed on the server, and AI selects only the single most important one to notify you about, or it may decide not to send anything at all, even if dozens of Events occurred.

With this approach, an app can be ranking in dozens of countries and moving gradually without generating any notifications. You’ll only hear about meaningful events, such as a jump of 50+ positions or entering the charts in a completely new country where the app has never ranked before.

You can also customize which types of Events you want to track and disable the ones you don’t need.

And if you’re tracking competitors, for example, you can turn off their notifications entirely and receive Push notifications only for your own apps.

@artsvit Artem, thank you for the feedback! I’m glad Toplify is useful for your apps. I built it primarily with indie developers in mind, since ASO services are often too expensive for them. That’s why I tried to make Toplify free for new and growing apps.

@busmark_w_nika Thanks for the idea! I’m actually planning to completely redesign the My Apps section in the next update, so it’ll be easier to separate your own apps from competitors and compare them more conveniently 🤟

@busmark_w_nika Definitely give the iOS version a try too. A lot of things are more convenient there than on Mac, and most people mainly use it for the convenience and mobility.

@arsen_kolyba Thank you for using Toplify since day one. It’s really great to hear that!

@silvia_odwyer1 Thank you for mentioning the interface! I put a lot of effort into making it simple and easy for everyone to understand.

Yes, I plan to expand the service in the future to track Google Play and Steam charts as well 🚀

@thisiskp_ Thanks for the feedback! As an indie developer, I initially built this service to solve my own problem, and then scaled it into a standalone product for everyone. I think it’s a must have for developers, especially those just starting out.

@shalabh_sanger Cool. Most developers do not always realize their app is ranking in some countries. Also, an iOS app can sometimes be ranking in the Mac charts, which many developers never check at all.

I’d be happy to hear your feedback once you try it. If your app has been in the charts at any point during the last two weeks, you’ll see it along with the exact dates and countries where it happened.

@yelyzaveta_kibets Well, both are true. But initially, we did limit it to two weeks because of the huge amount of data. Imagine storing every ranking change in every country for every app on the App Store. For example, a single game can have up to four separate rankings in each country: two category rankings, plus the overall ranking and the overall games ranking. We do plan to extend the history to one month once we grow a bit more 💪

@kate_ramakaieva Yes, unlike other similar services, you don’t even need to open the app. Just enable Ranking Alerts and go about your day. As soon as there’s a significant change in your app’s rankings, you’ll get a nice push notification that gives you a little mood boost.

@abrar_sami1 You’re welcome! If you have any questions or ideas for improvements, I’m always available.

@yelyzaveta_kibets Well, it’s hard to miss anything here because the AI simply chooses one notification to send, while all the other data is still there. You can open the app and see the full history for every country and category. At the moment, we store two weeks of history. So even if you add your app for the first time right now, you’ll immediately see its history for the past two weeks.

@domenicoflauto Thank you for the kind review. Yes, I 💯 agree with you. Every time I get a notification like this, it gives me a little dopamine boost and motivates me to keep improving my project. These notifications feel just as good as the ones from RevenueCat.

@andrewtanchuk Thank you for such a kind review of Toplify! I’m really glad you enjoy the app. I use your app, AppScout, and I think it’s the best alternative to App Store Connect and a must have for every app developer.

@nikitaeverywhere Thanks for the question! Web services didn’t really work for me for two main reasons: high pricing and the lack of real time ranking updates. They often provide daily or weekly reports, while I wanted to build a tool that shows the results of your work almost immediately, especially when you’re actively marketing your app.

For example, you post about your app on Threads and it takes off. You immediately get a notification that your app has entered the Top Charts. Then you generate a Share image to celebrate the result, post it, and potentially get even more downloads and purchases.

That simple feedback loop is what I built the product around.

@daryna_radiichuk Thank you! Yes, checking rankings manually across different countries is very time consuming and requires constantly switching App Store regions. Toplify takes care of all of this automatically, so there’s no need to check rankings manually or even open the app. Everything runs on the server.

@endore8 Thank you, Oleh, for supporting the project since day one! I’m really glad Toplify has been useful to you. I love your app Usage and have been using it for a long time. 💜

@ivan_4 When the app is in the Top 100. If it ranks in more than 4 countries, the limitations apply. For each app, you can see how many countries it’s currently charting in, along with a list of those countries.

@uaghazade If the salt rotates at midnight, a daily active user shows up as a brand new visitor every single day. That is a real tradeoff for privacy, but it also means any cohort built on returning visitors breaks completely rather than just getting fuzzier. How do you handle retention analysis without a stable ID across days?

@uaghazade The salted hash rotating at midnight is a clever privacy mechanism, but I'm curious about the analytics tradeoff it creates: a visitor who returns at 12:01am after visiting at 11:59pm would get a new hash and presumably count as a "new" visitor rather than a returning one. Is that an accepted tradeoff for the privacy guarantee, or is there a way to preserve session continuity across the rotation boundary?

@uaghazade awesome!

What pushed you to self-host + AGPL instead of just going full SaaS?

@thisiskp_ thank you very much for the name of our team. it's a pleasure to hear this from you!

@ilko_kacharov thank you a lot Ilko, for that kind of words!

@andrew_f3 The model never writes a query. Each MCP tool is a fixed, named read (site_overview, revenue_summary and so on) whose SQL is a constant on our side; the model only picks the tool and fills in parameters like the date range, and your chat client shows every call with those parameters. So what you audit is which read it chose, never hand-written SQL that could quietly redefine what a person is.

The layer above is pinned too, and it is yours: identity is only what you declare through identify(), scoped to your site, and answers that depend on it carry that scope with them rather than assuming it. On the revenue side there is no synthetic churn number to be fooled by: gross, refunds, disputes and fees stay printed as separate parts, so a refund you issued to yourself shows up as exactly that, a refund.

And for the front half of your story, the ignore flag keeps your own browser's visits out of the traffic numbers.

@yelyzaveta_kibets yes, that's the trade, and it's deliberate: anonymous ids rotate at UTC midnight, so we never track a cookieless visitor across days.

for your Monday-to-Thursday report: if your users log in, call identify(external_id) from the tracker. that hash is stable (still cookieless), and attribution then uses its full 30 day window, so the ad click and the purchase connect. purchases also get matched through order signals on the payment side, which often stitches the journey even without identify.

Either way the report shows which identity scope it was computed under, so it never claims a journey the data can't support.

@woshiliyana The "All clear" finish line is a nice detail most transcript editors leave you guessing whether you've actually caught everything. Curious about the flagging itself: is "uncertain" based on the ASR model's own confidence score, or something else (audio quality, cross-referencing multiple passes)? And does the flagged-word count change if I re-run transcription on the same clip, or is it deterministic?

@woshiliyana Great Launch Yan!

Was the audio-only-upload thing a privacy ask from users, or just how you'd build it anyway?

@woshiliyana Congratulations on Launch Yan.
The 30 subtitle styles caught my attention. Having that much choice while keeping the editing flow simple is a nice balance.

@galdayan It’s definitely on my radar. In an early two-speaker test, all four speaker turns were identified correctly, with confidence scores between 0.94 and 1.00, so the results look promising. Speaker labels and manual corrections are planned for v2. Overlapping speech is the harder case, and I want to test it on real interviews and podcasts before making any promises. Thanks for bringing this up!

@harkirat_singh3777 Thanks Harkirat

@heresalexandria Thanks, Alexandria! That’s exactly the direction I’ve been aiming for: bringing the useful parts of remote-control tools together and rebuilding the experience around modern agent workflows, rather than simply adding a chatbot on top. I also didn’t want Maccess to lock people into one “brain,” so letting people choose the AI tools they already trust is a big part of it. Really appreciate you noticing that!

@conduit_design Thanks André — good questions. It’s a hybrid: Maccess’s pairing, routing, and security layer is custom, while the underlying connectivity runs on managed infrastructure rather than a single self-hosted relay server. It prefers device-to-device connections where possible and uses a relay when network conditions require it.

The security boundary remains between your paired iPhone and Mac regardless of the route. Session content is end-to-end encrypted, control data uses ChaCha20‑Poly1305 authenticated encryption, and the relay never receives the decryption keys.

For local connections, yes—Maccess supports both the normal local network and Apple’s peer-to-peer Wi‑Fi path, which is the closest Apple-platform equivalent to Wi‑Fi Direct. That can get around router or client-isolation restrictions on some school and office networks. Availability is ultimately controlled by the OS and network policy, so it can’t bypass every managed environment; Online Access remains the fallback when local connectivity is blocked.

@galdayan Great question. On the same Wi‑Fi, the connection stays local. Online Access may connect directly or use a secure relay, depending on the networks involved.

Apple ID isn’t the only protection—it authenticates your Maccess account, while the iPhone must be separately paired using the QR code or pairing code displayed on the Mac.

Each session uses fresh encryption keys established with X25519. Control data—including trackpad input, keystrokes, clipboard data, and commands—is end-to-end encrypted and authenticated with ChaCha20‑Poly1305, so intercepted traffic is unreadable and tampered messages are rejected. Screen content is also encrypted between the paired devices.

If a relay is needed, it only forwards encrypted traffic and never receives the session keys. Maccess doesn’t store the content of your remote-control sessions, and you can disable Online Access or remove a paired device at any time.

@heresalexandria Great!

How'd you tune it to feel precise instead of laggy/twitchy like most webcam hand-tracking demos?

@heresalexandria The "camera never sends a frame anywhere" line is the right promise, and on-device is the only way this is usable at all. The number I'd want on the page before trying it as a daily driver: end-to-end
latency from hand movement to cursor movement, and at what frame rate. A webcam pipeline has a floor of roughly one capture frame plus inference (at 30 fps that's 33 ms before anything else) which is fine for gestures and rough for pointing. Have you measured it? Even a rough "about N ms on an M-series Mac" would make the pitch more credible, not less.

@yelyzaveta_kibets I love it! Thank you and of course, hope you have as much fun using it as I have :)

@yelyzaveta_kibets for sure, I genuinely appreciate an element of whimsy in novel applications of emerging technology!

So it adjusts "reach" automatically to account for being able to move the cursor around anywhere on your screen (and a bit beyond for scrolling/dragging) given the range of what your camera can see, but you can also adjust this manually if you find it isn't tuned quite right. When you genuinely do drop your hand out of frame, it'll release/stop controlling your cursor.

It sounds like the way your camera is facing, you might be able to benefit from using your phone's camera rather than webcam (a feature I just added thanks to @elias_iturri's comment in this thread), which would let you position it more ideally - the best results come from it looking head on at you from as close as you can get to the perspective of the monitor(s) you're looking at looking straight at you.

Please let me know if you hit any snags or have other questions :)

Excelsior,

Alexandria

@elias_iturri that's a great question! It didn't support iPhone camera, but thanks to your comment I've added it and tested - it works pretty well, although I'd advise disabling or loosening the "Only control while you face the screen" setting when using an iPhone or any other camera not attached to the screen(s) you're looking at:

Let me know if there's anything else you'd like to see in this or if you hit any snags.

Excelsior,

Alexandria

@galdayan absolutely, local was key not only for privacy but frankly speed for a feature you need running as quickly and efficiently as possible to reduce interaction friction.

Addressing false positives was the most fun part of the product development exercise here, ultimately addressed largely through ensuring gestures that perform actions are distinct as well as highly customizable and user-trainable.

By default, it only moves the cursor when you're looking directly the screen with an open upward facing palm in view (with adjustable sensitivity for all triggers). A click is registered by dipping just your pointer finger down towards your palm (as if clicking an invisible mouse in the air, but exaggerated) with the rest of the palm still open/splayed - a relatively unique motion.

All default tracking gestures were designed around the principle that they should start from the open palm and be fairly intentional movements - right click defaults to dipping your pinky in similarly to ensure maximum error boundary from other finger tracking, and scroll is triggered by dipping middle and ring finger in (sort of like you're a superhero shooting a web while scrolling websites).

Tracking is paused any time an open facing palm or face looking at the screen aren't detected (with configurability for use cases like broader angle of acceptability if you have to look across several external desktops to the left/right).

And finally, the fully user-trainable custom gestures (and out-of-the-box custom gestures like "shaka") can be as unique as your imagination allows, letting you bind motion that would conceivably never (or at least nearly never) occur without intention.

Let me know if that raises any other questions and hope you have fun using the app!

Excelsior,
Alexandria

@gitesh_chawda love that channels_first is supported natively via keras.config. does flipping that global flag mess with the numerical parity on the 13 segformer variants since upstream checkpoints are usually channels_last?

@galdayan caffeinate prevents the screen from dimming, but it doesn't prevent the sleeping when the lid gets closed. Amphetamine can get you there, but I found it quite junky to use. Just to note - all of this can be achieved via terminal commands, and I am in fact showing them in the FAQ section on the website.

I found that It gets annoying over time and I simply got tired of doing it over and over again. Nothing smart is going on with the sleep prevention implementation, it's just a more convenient way to flip the modes back and forth which I end up doing all the time.

@juansebsol Great Launch Sebastian!

What made "is this safe to stop" the feature to nail, over just listing ports?

@galdayan Good question. On unsupported Macs, the core port/process monitoring still works exactly the same and stays fully local. The Apple Intelligence-powered explanations are only available on supported Apple Silicon Macs, so on older Intel Macs that specific feature is gated off rather than silently sending anything to a cloud model.

I wanted “never leaves your machine” to be a hard rule, not something that changes depending on your hardware. I’m looking at adding a local heuristic process lookup fallback for unsupported Macs though, so they can still get useful context without sacrificing that.

@kotsutsumi The SSH/SFTP as regular panes design is the part I'd stress test hardest. A local pane's operations are effectively free, but a remote one has network latency built into every keystroke driven action, list, preview, copy. Does VeloFiler batch or prefetch to hide that, or does the vim style speed just get throttled down to how fast the SFTP connection responds on a slow server?

@kotsutsumi Built on GPUI like Zed, is Rust/GPUI as fast to build native macOS apps in as it feels to use? Would love to hear how that choice paid off (or bit you) during dev

haha

VeloFiler doesn’t store SSH passwords or private keys. It uses your OpenSSH configuration with SSH-agent or key-based authentication, while macOS permissions control access to the relevant SSH files. Your credentials stay on your Mac and aren’t sent to our servers.

Thanks! VeloFiler asks for access when you choose a folder, then uses macOS security-scoped bookmarks so you don’t get prompted for every navigation. The App Store route was definitely harder, but it was worth it for a trusted native install.

@omarships Great Launch!

What made you pick "one win, ever" instead of a recurring daily leaderboard reset?

@omarships I’m intrigued by the deterministic replay mechanic, and how does the seeded simulation ensure fairness across different devices and internet speeds?

@alexcloudstar the "without crossing lines" line is the part I'd want to understand better before trying it. nothing technically stops someone from dropping a handle that isn't their own, right? the fun framing works fine when it's your own account you're paying to get roasted, but a public leaderboard card that anyone can point at anyone else's handle is a different product. is there any check that the person submitting is roasting themselves, or is it fully open by design

@awt the few-days comment delay is an interesting way to protect the community from random negativity.
curious if you’ve already seen a difference in the quality of feedback because of it?

@awt Great Product Andrew!

Is that private craft score AI-generated?
Also, Curious how you keep people from writing for the score instead of the poem. haha!

@awt the "authenticity" check is the part I'm most curious about, separate from the craft score question below. if someone pastes an AI-generated poem and claims it as their own, is there an actual detection step, or is it more of an honor-system flag that gets sorted out if readers or other poets call it out after the fact? seems like the hardest thing to enforce given how good AI poetry can sound on a surface read

@david_pfluegl1 the power button caveat you flagged is the interesting one to me. on Apple Silicon, holding it long enough gets you into recovery mode, which in theory could be a path around the whole lock for someone determined enough at a trade show booth. is that actually possible in practice, or does something like the exit password/FileVault end up blocking that too, just not by design

@ben_shih Awesome Launch Ben!

On the 73 motions, any thought of a "surprise me" shuffle for people who don't want to scroll?