Stringers
Mobile App · iOS · written 11-14
SkyCast · Hefei Yunluo Technology Co., Ltd.

A free sunset forecast went offline for a month, and a forecaster with 40k followers launched a ¥139/year app

What it is A Chinese app focused on sunrise and sunset glow (cloud color) forecasts, with short-range nowcasts, an alarm that only rings when the glow hits your threshold, and a 24-hour satellite cloud view. Monthly ¥14, annual ¥139, China store only.

Why it took off The author was already a sunset forecaster with 40k followers on Xiaohongshu, and built a base first through free websites, a mini program, and sponsorships. Then sunsetbot went down, leaving people chasing sunsets looking for a replacement, and the weather category's low bar let a few dozen subscriptions put it on the chart.

CHow you could follow this Don't copy a sunset app; take the path: first build trust as a free forecaster in a niche you know (tides, auroras, snow conditions, stargazing) where people "live by the weather," then launch a paid tool with threshold alerts, or partner with forecasters in Japan, Taiwan, or Korea to bring the model there.

C grade: its rise came from the author's year-plus of built trust and a one-off gap from sunsetbot's collapse, which followers can't replicate. The domestic sunset niche is already crowded with quick-built tools, and the paying pool is small. What's worth learning is the "expert blogger turns into a subscription tool" path; copying it directly isn't worth it. Few people know about it yet

Est. monthly revenue (App Store) under $1KWe estimated this from its chart ranks and rating growth across storefronts, calibrated against verified revenue; typical error is 2–3x. Read it as a tier, not an exact number.

Data (updated daily) · First launch 2026-11-10 · China ratings 1 · Charted in 11-20 of 2 stores
iOS · best category chart rank · by store
10020011-1211-1411-1611-1811-2059 Hong Kong72 China
ChinaHong Kong
Data 2026-11-12 → 2026-11-20 (9 days, 9 observations). Lower rank is higher; gaps mean it wasn't charted that day
iOS · Rating count · China
00.5111-1211-1411-1611-1811-201
Data 2026-11-12 → 2026-11-20 (9 days, 9 observations). Apple doesn't publish download counts; rating growth is a proxy for downloads.

SkyCast is a Chinese app focused on sunrise and sunset glow forecasts. It launched in Nov 10 and hit #42 on the China store's weather Top Grossing chart the next day. Its creator has posted free sunset forecasts twice a day on Xiaohongshu for over a year. At launch, sunsetbot, the free tool most domestic sunset chasers used, had already been down for nearly a month. This piece covers what it rode on, what the #42 rank is worth, who is competing, and what you can take from it.

"The sunsets I chased for three years: sunsetbot has really shut down"

In Nov 15, a Hangzhou user called "快门嘎嘣脆" posted on Xiaohongshu: "I remember the first time I learned about this site, I refreshed it at 1 a.m. to see a predicted big morning glow, trusted it once, and actually got the shot when I went out... In summer sunset season I got into the habit of checking it three times a day." The post ends: "There may be no reliable software left to use; chasing sunsets is going to be more a matter of luck."

sunsetbot was a free website and mini program built by a Shanghai Jiao Tong University student over 2022 years, giving morning and evening glow forecasts by city, and it was the most-used tool in the domestic sunset-chasing community. From mid-October onward it stopped showing data. On Oct 13, someone asked on Xiaohongshu "sunsetbot's server has been down for days, what other channels are there?" It got 64 likes and 26 comments, and one reply said: "The cloud-map program it relied on shut down... now you basically can only use Lijing Weather, and the rest aren't accurate." Over the following weeks, posts like "what's up with sunsetbot" and "without sunsetbot, everyone's getting it wrong" appeared every few days.

On Oct 30, a Xiaohongshu forecaster called "千新星" posted a "SkyCast beta tester recruitment," prioritizing fans who "follow forecasts long-term and occasionally send back photos or feedback." On Nov 10, the app launched.

What it is

Open SkyCast and you get a national map of sunset-glow forecasts, scored and graded, telling you where the sky may "burn" tomorrow morning or tonight. It has a few things beyond a typical forecast:

  • Nowcasting: it recomputes clouds and aerosols for the next few hours using satellite and ground observations, which the author says is noticeably more accurate than ordinary numerical models;
  • Alarm only when the threshold is hit: the alarm rings only when the glow forecast reaches your set threshold, because the worst thing for sunset chasers is waking at five a.m. and finding no glow;
  • 24-hour visible-light cloud imagery: you can see clouds at night too, plus 20 layers including wind, precipitation, aerosols, and light pollution;
  • Astronomy data such as the sun's position for sun-halo events, stargazing indices, meteor showers, and auroras.

Simplified Chinese only, China store only. Free to download; Pro membership ¥14/month, ¥39/quarter, or ¥139/year; new users get 3 days of membership free (the China App Store page and the author's launch post, 11-14). The developer is Hefei Yunluo Technology, registered 2026 years ago, and this is its only app.

The author's Xiaohongshu bio reads "theoretical physics | atmospheric science | astrophysics," with 39.9k followers, 81k total likes, and an Anhui IP. The forecasts are not new: back in September 2025 he wrote on his own site astsky.com that "the past year's forecasts were delivered as updated forecast posts" and "the forecast service is currently free, supported by ads." Later he added a mini program and two sponsorship tiers on Afdian, at ¥5 and ¥8 per month. The app is the paid version of all of this.

What a #42 ranking is worth

DateWeather free chartWeather Top Grossing chart
11-123942
11-134257
11-144976
11-155942

(App Store China, Weather category)

As of Nov 15, it had only 1 ratings. The free chart was slowly sliding, and the grossing chart returned to #42 on the fourth day: fewer downloads, but people are still paying.

The bar for the China weather Top Grossing chart is low. On the same day, Solora, with 313 ratings, ranked #64, and the long-established overseas app Alpenglow ranked #80; Lijing Weather, which has spent five or six years on sunset and photography weather in China (786 ratings), ranked #17. At that level, a few dozen subscriptions a day would put an app in the top 50 (inferred), though revenue can't be verified. So #42 shows that the payments are real, not that the volume is large.

Where the users come from: a person posting two forecasts a day

The author's own Xiaohongshu account is the main channel. 千新星 posts two items daily, "tomorrow's sunrise glow forecast" and "today's sunset glow forecast," each getting a dozen to thirty-some likes; other bloggers who repost his forecasts credit "data from 千新星." His pinned explainer "A chart to understand sunset glow and common misconceptions" has 1,495 likes. The Nov 10 launch post got 148 likes and 49 comments, and he later posted an introduction to nowcasting and the Android release (Xiaohongshu, 11-14).

Old users moving over. The launch post said: "If you've often given feedback on forecast results, or sponsored the early forecasts, DM me for a longer membership." The people who built up during the website and mini-program period, and Afdian sponsors, were the first paying users.

The gap left by sunsetbot. From mid-October to November, people on Xiaohongshu kept publicly asking "what can I still use?" SkyCast's beta recruitment and launch both landed in that gap. There's no direct data on how many former sunsetbot users actually switched.

不是靠商店搜索。 在中国区 App Store 搜「火烧云」,前十里没有 SkyCast(11-14),排在前面的是 Solora、莉景、天文通,还有 10 月 20 日刚上的「霞 · 朝霞晚霞火烧云预测」。作者在上架帖里让粉丝「应用商店搜索 SkyCast,认准图标」,评论区就有人下错了:「是这个吗?我怎么全英文看不懂」,作者回「下错了,看看封面图的 App 图标,一个无穷符号」。下载基本是粉丝拿着名字去搜的。

No sign of it on Douyin, Weibo (千新星's Weibo has only 237 followers), Bilibili, or X, and no media coverage.

How longtime fans view the paywall

Most comments under the launch post were "already downloaded," "downloaded right away," "have to support, let's push the download numbers up," and "next year when I'm chasing sunsets I won't have to ask you here." There were also dissenting voices:

  • "A subscription is really unfriendly for office workers; I barely have time to watch sunsets to begin with."
  • "Teacher 千, could you offer a one-time purchase for longtime fans?" The author replied: "No plans for a one-time purchase right now, since important features will keep coming."
  • "I remember another app had a payment option where, on the day you want it, you just pay a few bucks for a one-day membership." The author replied: "Right now the app's monthly subscription averages less than ¥0.5 a day."

Chasing sunsets is a low-frequency, seasonal thing. One Hangzhou user said that after autumn brings more haze, sunset-chasing became a "summer-only treat." There's a gap between a subscription model and that usage pattern, and SkyCast hasn't solved it yet.

Who's competing

ProductFormatScaleHow it differs from SkyCast
sunsetbotWeb + mini program, from 2022Used to be the most-used free tool in the circle; went down in mid-OctoberFree, forecasts by city
Lijing WeatheriOS + mini program, from 2020786 ratings (average 2.4), #17 in Weather Top GrossingVeteran app with broad features (sea of clouds, rainbows, photography tools), reputation is mixed
TianwentongiOS, from 2019963 ratings, #11 in Weather Top FreeLeans toward stargazing; sunset glow is a side feature
SoloraiOS, from 2024313 ratings, #64 in Weather Top GrossingForeign developer, name stuffed with Chinese keywords to grab search traffic
Xia, Sunset-chasing Plan, LOHUI sunset-glow map, a few open-source toolsiOS / web, around 2026-10from 0 to a few hundred likesQuickly built by people after sunsetbot went down; mostly public-model data plus scoring

One more point: some companies already sell ready-made "fire-cloud forecast APIs" that anyone can plug into their own weather app. So a fire-cloud index by itself isn't a moat. SkyCast's edge is, first, its own nowcasting and cloud imagery, and second, the trust the author built by publishing forecasts publicly every day for over a year and checking them against observations. The second can't be bought.

If it were you

Bottom line: SkyCast can't be copied easily. It took off because the author built trust over a year-plus and a group of people who check his forecasts every day, plus a one-time gap from sunsetbot going down. In China's fire-cloud niche, a batch of quickly built tools has already moved in since October. What's worth taking is the path: start as a free expert creator, then turn the forecasts into a subscription tool.

Who it suits: People who already have real expertise in a "weather-dependent" niche and a content account: fishing with tides, auroras, sea of clouds, skiing with snow conditions, stargazing. Or developers who can partner with creators like this.

Who it doesn't suit: Indie developers with no following who want to build a "more accurate fire-cloud app" and fight for search traffic in the store. Algorithms can be bought; trust can't. Also, paying users in this niche are a small pool, and a #42 grossing rank may only mean a few dozen orders a day.

How to do it:

  • Publish forecasts first, then build the tool. Post every day, check against observations, and show your hits and misses openly. Build trust over a few months, and cover costs early with ads and sponsorships.
  • Charge for "watching it for you": target alerts and nowcasting. Basic forecasts stay free.
  • Fill the need SkyCast doesn't cover. Comments show people want a one-time purchase, or want to "pay a few bucks to use it for a day." In a niche with many low-frequency users, per-day, per-use, or one-time pricing may fit better than subscription only.
  • Do the same thing in another market. Searching "夕焼け 予報" on the Japan App Store, none of the top ten are dedicated apps with real traction. The newest one launched in Nov 6 with 0 ratings (Japan App Store, 11-14). I haven't verified whether big weather apps there already include this kind of forecast. Taiwan and South Korea work the same way, provided you can find local forecast creators.

How to get users: The creator's own account is the main channel. Recruit early testers from existing fans first. Give longer memberships to past sponsors. When a free tool in the circle goes down, show up under the "what can I still use?" posts.

When to move: For China's fire-cloud niche, winter isn't peak sunset-chasing season. The next window is before next spring and summer's sunset season. Whoever builds an accuracy reputation then and picks up users from sunsetbot wins. Switching circles or countries has no window like this.

What would show I'm wrong: If SkyCast reaches the top 20 in Weather Top Grossing during next summer's sunset season, the niche's paying pool is bigger than estimated here, and it's worth going earlier and harder. If a quickly built tool with no creator backing also makes the top 50, distribution matters less than I think, and copying the approach works. Conversely, if sunsetbot comes back, or Lijing adds nowcasting and target alerts, the gap in China closes.

iOS store page ↗

Research notes: competitor table, acquisition evidence, full research report
Competitors
NameWhereLaunchedScaleHow it differs from this one
sunsetbot (fire-cloud analysis and records)Web2022The free website and mini program most used in China's sunset-chasing circle; 2026-10 mid-period it stopped producing data, and 11-15 posts say it has shut downFree, cross-section charts by city. SkyCast offers a nationwide grid map, nowcasting, and target alerts for a fee.
Lijing WeatheriOS2020-07 (mini program 2019-08)iOS CN 786 ratings, average 2.4, #6 in Weather Top Free and #17 in Weather Top GrossingVeteran app with broad features (sea of clouds, rainbows, photography). SkyCast focuses only on fire clouds, with its own nowcasting assimilation system and the author's personal reputation.
TianwentongiOS2019-09iOS CN 963 ratings, #11 in Weather Top FreeLeans toward stargazing and astronomy events. SkyCast focuses on fire clouds; astronomy is a side feature.
Solora (fire clouds, sunset, sunrise)iOS2024-08CN 313 ratings, #64 in Weather Top GrossingForeign developer, name stuffed with Chinese keywords, uses a generic model. SkyCast has domestic nowcast correction and creator-based distribution.
AlpenglowiOS2016US 2,144 ratings, CN 48Long-running overseas app with few Chinese users and no nowcasting
Xia: morning and evening glow and fire-cloud forecastsiOS2026-10-200 ratings, #175 in Weather Top FreeA personal project with simple forecasts, no proprietary data and no existing following
Sunset-chasing Plan (Skygrow), LOHUI and similar web toolsWeb2026-10Single Xiaohongshu posts get tens to a couple hundred likesQuickly built by AI using public-model scoring. SkyCast has proprietary cloud imagery and an assimilation system.
Xingtu Earth fire-cloud forecast APIOtherLiveB2B interfaceAnyone can plug it into their own weather app, which shows a fire-cloud index by itself isn't a moat
Where users come from
  • Verified the author's own Xiaohongshu account (the KOL is the founder): 千新星, 39.9K followers, IP in Anhui, posts two sunrise/sunset forecasts a day; 11-10 launch post with 148 likes, 64 saves, 49 comments; 11-12 nowcast post, 11-14 Android post
  • Verified Pre-launch follower beta recruitment: 10-30 Xiaohongshu and Toutiao "Xingqiong Telescope" simultaneously posted "SkyCast Beta Recruitment," prioritizing fans who check forecasts long-term
  • Verified Migration of existing users (website, mini program, Afadian sponsorship): astsky.com 2025-09 already offers free fire-cloud forecasts, funded by ads and Afadian sponsorships at ¥5/¥August; there was a mini program before; the launch post says early feedback givers and sponsors can DM for longer memberships
  • Verified Platform event: sunsetbot outage: From 10-11 to 11-15, multiple Xiaohongshu posts said "sunsetbot is down," "everything went wrong without sunsetbot," "it's really shut down," and users publicly looked for alternatives. SkyCast 10-30 recruited beta testers and 11-10 launched.
  • Inferred Store search: The author told fans to "search SkyCast in the App Store and look for the logo," and some people downloaded a same-named app by mistake. I didn't see any keyword optimization.
  • Inferred Ads, Douyin, Weibo, media: No matching Douyin account. Qian Xinxing has only 237 followers on Weibo, x-radar has 0 posts, and there's no media coverage. Not covered by the ad library.

SkyCast (iOS 6825403466) research report

In one line: SkyCast is a subscription app built from the forecast algorithm of indie developer Qian Xinxing, who has run a "daily fire-cloud forecast" on Xiaohongshu for a year or two (39.9k followers on Xiaohongshu, IP location Anhui, matching the developer's company Hefei Yunluo Technology). It launched 11-10 and reached #42 in China's Weather Top Grossing the next day. It worked for three reasons: the author is already a sunset-forecast creator in the chasing community, with a built-in group of fans who check his forecasts every day; sunsetbot, the most-used free fire-cloud tool in China, has been down since 2026 mid-October, and sunset-chasers are looking for replacements; and the bar for China's Weather Top Grossing is low, so a few dozen paying users can get on the chart.

1. How to read the data card

  • Listing and versions: Listed 2026-11-10, first version 11-11 1.0.0; Simplified Chinese only, China store only (also visible in the HK store, where it once charted in Top Free); free download with in-app purchases (iTunes lookup, cn, 11-14).
  • In-app purchases: Pro annual ¥139, quarterly ¥39, monthly ¥14, single month ¥16 (App Store cn page, 11-14). New users get 3 days of membership on sign-up. The app gives this itself, it is not an Apple free trial (the author's Xiaohongshu launch post 11-10).
  • Charts (appstore db chart_rank, cn, weather category 6001): | Date | Top Free | Top Grossing | |---|---|---| | 11-12 | 39 | 42 | | 11-13 | 42 | 57 | | 11-14 | 49 | 76 | | 11-15 | 59 | 42 | HK weather Top Free 11-14 #128, 11-15 #194. Top Free is slowly sliding, Top Grossing 11-15 climbed back to 42.
  • Ratings: 11-15 ratings in total up to 1 (5 stars), no reviews (appstore db app_day). Near-zero ratings are because it has only been live a few days with a very small user base, not because nobody uses it.
  • What it takes to hit Top Grossing #42: the bar on China-store weather Top Grossing is low. On the same day (11-15), Solora (313 ratings) was #64 and Alpenglow #80; 莉景天气 (786 ratings, the veteran domestic fire-cloud/photography weather app) was #17 (appstore db app_profile). Judging from this, a few dozen subscriptions a day is enough to reach the top 50. I have not verified the actual revenue.

2. Competitors

CompetitorsPlatformLaunchedScaleHow it differs from SkyCast
sunsetbot (fire-cloud analysis and records)Web sunsetbot.top + WeChat mini program + WeiboSince 2022 (built by a Shanghai Jiao Tong University student, covered in SJTU news)The free tool most used by the China sunset-chasing community, with many Xiaohongshu posts mentioning it. It has not been able to show data since mid-2026-10, and by 11-15 posts were saying "sunsetbot really shut down".Free, gives atmospheric cross-section maps by city. SkyCast offers a nationwide grid map, nowcasting and alerts, and it charges.
Lijing WeatheriOS + WeChat mini programMini program 2019-08, iOS 2020-07iOS cn 786 ratings (average 2.4), weather Top Free #6, Top Grossing #17; still being updated as of 11-10The veteran. Sunrise/sunset glow, cloud sea and rainbow forecasts plus photography tools, with many users; reviews are mixed. After sunsetbot went down, Xiaohongshu comments said "now you can basically only use 莉景".
TianwentongiOS2019-09iOS CN 963 ratings, #11 in Weather Top FreeLeans toward stargazing and astronomy, also covers fire clouds. SkyCast focuses on fire clouds, with astronomy as an extra.
Solora (fire clouds, sunset, sunrise)iOS2024-08CN 313 ratings, #64 in Weather Top GrossingBuilt by a foreign developer who stuffed Chinese keywords into the name to win search. Uses generic models with no domestic nowcasting correction.
AlpenglowiOS/Android2016US 2,144 ratings, CN 48The oldest sunset-prediction app overseas, few Chinese users.
Sunsethue / Helio / PhotoPillsiOS2026-07 / 2020 / 2013Both small (PhotoPills is a paid photography tool).Overseas, weak Chinese support.
Xia: morning and evening glow and fire-cloud forecastsiOS2026-10-200 ratings, #175 in Weather Top FreePersonal apps that popped up after sunsetbot went down.
Sunsetter Skygrow, LOHUI sunset map, will-sky-burn-tonight, couletian's open-source forecast tool, Yeats's forecast site, r-ayin/sunset-predictionWeb / open sourceAround 2026-10Individual Xiaohongshu posts get dozens to a couple hundred likes; couletian's post got 129 likes and 144 saves.All built with AI by people after sunsetbot went down, mostly on public model data with simple scoring.
Xingtu Earth's "fire cloud forecast API".B2B interfaceLive—Anyone can buy it and plug it into their own weather app, which shows the "fire cloud index" itself is not much of a moat.
Big weather apps like Moji and ColorfulClouds.Big-company features—The top few spots on the weather Top Grossing chart.No dedicated fire-cloud/nowcasting product (not checked one by one).

Verdict: this niche has been built many times over and is not short of competitors. But in China it has always run on free tools (sunsetbot, mini programs) plus human forecasts from bloggers. SkyCast is the new one that focuses on fire clouds, accepts subscriptions, and has its own data assimilation system.

3. Where users come from

  1. The author's own Xiaohongshu account (confirmed, main channel). The profile for "千新星" reads "Theoretical physics | Atmospheric science | Astrophysics 🔹SkyCast", 39.9k followers, 81k likes and saves 12k, IP location Anhui (TikHub Xiaohongshu get_user_info, 11-14). The pinned science post "Understanding the principles and common misconceptions of fire clouds in one chart" (2025-08) has 1,495 likes and 985 saves. He posts two items a day, "tomorrow's sunrise glow forecast / today's sunset glow forecast", each getting a dozen to thirty-odd likes, with up to 26 comments. 11-10 posted "SkyCast is now live on the Apple App Store" (148 likes, 64 saves, 49 comments), 11-12 posted an introduction to the nowcasting feature, and 11-14 announced the Android launch. Other bloggers also credit "data from 千新星" (北海の天气之子, 9-11), and 11-12 has a user writing "forecast and data source SkyCast".
  2. Pre-launch beta recruitment (confirmed). 10-30 posted "SkyCast Beta Recruitment" on Xiaohongshu and Toutiao's "星穹望远镜" at the same time, prioritizing followers who "follow the host's forecasts long-term and occasionally share photos or give feedback" (57 likes on Xiaohongshu). 10-13 first posted a preview of the astronomy page, saying the app was "going through filing".
  3. Migration of existing users (confirmed). These forecasts already existed: on astsky.com, 2025-09 has an article explaining the fire-cloud forecast (credited to "千新星") saying "the past year's forecasts were delivered by updating forecast posts" and "the forecast service is currently free, operated on advertising". There was also a WeChat mini program before (10-07, 11-02 posts both mention it), and afdian.com/a/skycast has two sponsorship tiers at ¥5 / ¥8 a month. The launch post says people who often gave feedback or sponsored early can message for a longer membership. So the plan was to build an audience with a free website, a mini program and sponsorships, then move them to a subscription app.
  4. Platform event: sunsetbot went down (confirmed). On Xiaohongshu: 10-11 "Is it just me who can't log into sunsetbot?", 10-13 "sunsetbot's servers have been down for days, what other channels are there?" (64 likes, 26 comments), 10-17 "What's wrong with sunsetbot?", 10-25 "Has Sunsetbot shut down?", 10-30 "Without sunsetbot, everything's broken", 11-15 "Three years of chasing sunsets, sunsetbot really went under". Some comments say a cloud-image program it depended on was shut down. SkyCast happened to be recruiting beta testers at 10-30 and launched at 11-10. How many sunsetbot users actually switched over is not directly known.
  5. App store search (inferred). The name SkyCast is a generic English phrase, so searching "火烧云" in the cn store may not rank it near the top. The launch post asks fans to "search SkyCast in the App Store and look for the icon", and some commenters downloaded a different English-named app called SkyCast by mistake. So right now fans carry the name over and search for it, rather than organic keyword traffic bringing people in.
  6. Not found: Meta/TikTok ads (the ad library only tracks listed products, so not finding ads doesn't mean none were run), Douyin (no matching account for "千新星" or "SkyCast"), Weibo ("千新星" has only 237 followers), Bilibili, X (x-radar shows 30 posts in the last 0 days), and no media coverage.

4. Why it took off

  • Built-in targeted users: the author spent one or two years as a free forecaster in the sunset-chasing community. His 40k Xiaohongshu followers are mostly landscape photographers and sunset chasers who decide whether to go out based on his forecast each day. Once the app launched, that group had a reason to pay.
  • Timing: sunsetbot went down from mid-October, and sunset chasers publicly asked on Xiaohongshu "what else can I use?". SkyCast launched right in that gap and picked up the demand the free tool left behind.
  • The product is real: it is not a generic model with a skin on top. It has an in-house 24-hour visible-light cloud imagery (you can see clouds at night), Aether data assimilation and nowcasting (the typhoon 3D cloud case is on the website), alerts that wake you when the sunrise glow hits a threshold (solving the "set an alarm, get up, and nothing burned" problem), and multiple models (EC/GFS/ICON) with correction. In a niche mostly made of "generic model + scoring", this level of technology is rare.
  • The price fits: ¥14 a month, ¥139 a year. The author said in comments that "a continuous monthly plan works out to under five jiao a day". That's not expensive for people who can afford a camera and will drive to chase a sunset.
  • The chart bar is low: the top 50 of China-store weather Top Grossing needs only dozens of orders a day. So reaching #42 shows conversion to paid is real, but the volume is small.

5. Differentiation

Compared with sunsetbot, 莉景, and Solora, SkyCast is different because of "nowcasting + alerts that wake you when conditions hit" combined with in-house cloud imagery and assimilation, backed by the author's own atmospheric physics background and the credibility built by publicly checking his forecasts against reality for a long time. On Xiaohongshu he posts forecasts daily and then reviews them against observations (e.g. "9.20 Shanghai sunrise glow nowcasting performance"). That credibility is something a tool competitor would struggle to buy.

This difference is only part of the story. The more direct reasons are distribution through "the author is the forecast blogger" and sunsetbot going down. The technology determines whether the app can keep paying users, but what got people to download it for the first time was the author's account and that gap.

6. Visibility and copycats

  • Visibility: still not many people know about it. Revenue isn't public, and no media outlet or large X account has written about it; it spreads only within the author's Xiaohongshu fan base and Toutiao. The app has only 1 ratings.
  • Copycats: nobody is directly copying SkyCast yet. But the gap sunsetbot left has already drawn a wave of individual developers (Skygrow, LOHUI, 霞·App, open-source tools, etc.), and most were built quickly with AI around 2026-10. This niche is going to get crowded.

7. What others can learn

  • Be a free expert blogger first, then sell a tool: publish forecasts publicly every day, build credibility by reviewing them against observations, support yourself with ads and Afdian sponsorships at first, and finally convert followers to subscriptions. Any small niche with willingness to pay can do this, such as tidal fishing, aurora, cloud sea, birdwatching, and ski conditions.
  • Watch for the moment free infrastructure falls over: when a free tool everyone in a community uses goes down, users publicly look for replacements on Xiaohongshu, and that is the cheapest moment to launch and show up under those posts.
  • Use "wake me / remind me" as the paid hook: "alert only when conditions hit" turns a forecast from "something you glance at" into "something that watches for you", which makes renewals more likely.
  • Copy it into other languages: overseas there are Alpenglow and Sunsethue, but the "nowcasting + threshold alerts + blogger distribution" combination can be repeated in Japan (夕焼け予報), Taiwan, and Korea.

8. Still unclear

  • Actual downloads and revenue: I can't verify how much money corresponds to the #42 Top Grossing ranking. Ratings are too few to show download growth.
  • There is no direct data on how many of sunsetbot's old users actually switched over.
  • Whether the Toutiao account "星穹望远镜" belongs to the author: the content matches Xiaohongshu, but the account also posts telescope buying guides, so it may be a partner or repost account.
  • The Android version is currently only available from the website (11-14). What happens once it is on the app stores is not known yet.
  • Whether users stick around: the fire-cloud season in winter differs from the photo-worthy opportunities, so whether subscriptions keep renewing needs a few more months to judge.
  • Compute costs: the author repeatedly mentions "insufficient compute", and nowcasting is expensive to run. There is no data yet on whether subscription revenue can cover it.

Sources

This article was written by AI from public data and the data we collect daily. Numbers are sourced, and inferences are labeled.