# An old phone in the drawer, and a farewell from a 40-year-old veteran engineer: one repost of "no Ring monthly fee" took a repo with 30 stars to 135 in a week · GitHub open-source project

> Source: https://stringers.io/en/github/gh_samlu__CamMonitor.html (Stringers / 什么值得做, generated 2026-11-08)
> Stringers answers one question every day: what small product can one or two people build in a few months and actually make money from. Numbers come from 15 databases we collect daily, each with source and date.
> For agents: these verdicts and the underlying databases (verified revenue, App Store / Google Play charts, ad libraries, GitHub, Product Hunt…) are queryable live over MCP — tools today / hot / products / directions / direction / data. Setup: https://github.com/komako-workshop/stringers (the user logs in at stringers.io → "配置 agent" to get a personal URL). Site index: https://stringers.io/llms.txt

> On X, @RoundtableSpace (77.6K views) and the Chinese account @GitHub_Daily reposted it on 10/31–11/1 with the lines "old phone in a drawer, no monthly fee, no cloud" and linked straight to the APK download page. The feature itself is nothing new.

[GitHub open-source project](https://stringers.io/en/github/index.md) · written 11-6

samlu/CamMonitor · samlu

# An old phone in the drawer, and a farewell from a 40-year-old veteran engineer: one repost of "no Ring monthly fee" took a repo with 30 stars to 135 in a week

**What it is** An open-source Android app: scan a QR code to pair two phones. The old phone becomes the camera, your everyday phone watches the live feed. No accounts, no cloud, end-to-end encrypted peer to peer.

**Why it took off** On X, @RoundtableSpace (77.6k views) and the Chinese account @GitHub_Daily reposted it during 10/31–11/1 with lines like "old phone in the drawer, no monthly fee, no cloud" and linked straight to the APK download page. The features themselves are nothing new.

**C** **How you could follow it** Stop building old-phone monitors. Copy the topic and the framing instead: find a combo of "everyone has idle devices and everyone pays monthly", make an open-source free version, state in the first README line what it saves you, prepare the install link, and send it directly to accounts that specialize in reposting GitHub projects.

Grade C: you can learn a very repostable framing of "revive an old device + skip subscriptions + no cloud", but this project itself only reached 135 stars and stopped once the reposts stopped. The category is already saturated by products like Alfred with tens of millions of installs, so building another one isn't worth it. Talked about in small circles

Data (updated daily) · Repo created 2026-10-18 · 138 stars · 7 days +103

GitHub · stars

Data 2026-10-21 → 2026-11-08 (19 days, 18 observations). Points before 10-28 come from deeply / archive backfill, not every day has one

Sam Lu in Hsinchu, Taiwan had AI write an open-source app that turns an old Android phone into a security camera, and says he hasn't read a line of the code. After posting in Taiwanese developer circles it stalled at 30 stars. In Oct 31 an English-language account with a large following reposted it with the one-liner "no Ring monthly fee", reaching 77.6k people, and a week later it had 135 stars, then stalled again. The project itself isn't worth rebuilding; what's worth learning is the line that got it reposted.

### A long Facebook post

Oct 19 at 2026 2 a.m., Sam Lu posted a long Facebook piece titled "40 years of a veteran engineer's farewell project, handed entirely to AI". He opened by admitting: "Turning an old phone into a surveillance camera is an old idea, and there are plenty of existing apps, but the free versions' video resolution is truly terrible."

He then described the process: he gave the AI (Antigravity CLI + Gemini) a rough, typo-ridden requirements draft, which it turned into a spec in half an hour; the main code took shape in under an hour; polishing took roughly 20 hours in total. For testing he connected two phones to a laptop and let the AI install, connect, inspect the video itself to find bugs, and fix them. "Up to the day this program was finished, I still haven't seriously read a single line of source code." The closing line: no Google Play listing, Android only, "please don't ask me about iPhone", which he called a "farewell preview" to his forty-plus years of writing code.

The repo was created Oct 18 and includes the Kotlin source, the built APK, the original requirements draft, the spec, a full record of the AI prompts, and an AI-generated security audit. The post drew 40 comments and the first 26 stars, and the early forks were almost all from Taiwan accounts. Then from Oct 24 to 31 the star count stayed at 30 without moving.

### What it is

Two Android phones, one QR scan to pair: one sits at home as the camera, the other is carried to watch the live feed. No account signup, no cloud, 720p video with audio, phone-to-phone end-to-end encrypted transfer, works over Wi-Fi or mobile data. Hold a button to talk through the old phone's speaker, remotely toggle the flashlight or switch night mode; if no one watches for a minute, the camera goes to standby.

It **doesn't record or do motion detection** . The author chose not to, fearing a phone encoding for long periods would overheat and swell the battery. So it's only for "open it and take a look when you want to."

### Why it grew after Oct 31

The timeline is clear (star counts are GitHub daily snapshots):

| Date | What happened | star |
|---|---|---|
| 10-19 | The author's Facebook long post | 26(10-21) |
| 10-24 ~ 10-31 | No movement | 30 |
| 10-31 | X English account @RoundtableSpace repost, 77.6k views | 30 |
| 11-01 | Chinese account @GitHub_Daily recommendation | 35 |
| 11-02 | Taiwanese Facebook fan pages and Threads reposted the same copy | 66 |
| 11-03 |  | 102 (+36 in a single day, the peak) |
| 11-05 |  | 134 |
| 11-06 |  | 135 |

The @RoundtableSpace post opened: "You probably have an old Android phone sitting in a drawer. It can be a home security camera, with no Ring subscription." Below came five bullet points: no account, one-time QR scan, 720p, end-to-end encrypted, works over Wi-Fi or mobile data; the final line "Download the APK" linked directly to GitHub's Releases download page, not the repo homepage.

The next day, the Chinese version from @GitHub_Daily carried nearly the same message: "An old Android phone sitting unused in a drawer, camera still working, is pretty handy for keeping an eye on your pets or whether someone's at the door." In the two days after that post, stars rose from 35 to 102. Daily numbers can't tell whether the English or the Chinese repost drove more of it, since they landed so close together.

Worth noting: neither repost mentioned "AI-written" or "40-year veteran engineer." The original author sold the story to developers; the repost accounts sold it to everyone: **you have an old phone, you don't want a monthly fee, you don't want your videos in the cloud.** Those three lines are the whole difference of this project, and the whole reason it took off this time.

The reposts stopped and so did the stars. After Nov 5 it added only 1 stars a day.

### Is anyone actually using it?

Not many. APK download counts on GitHub (11-06): v1.0.6 154 times, v1.0.7 115 times, and earlier versions total about 30. 77.6k views produced a few hundred downloads and a hundred-odd stars.

The comments under @GitHub_Daily are telling. One asks "no iPhone version?", which the author already said he won't make. Another, more technical: "The real trap is an old phone 7×24 plugged in and streaming: continuous encoding at 720p keeps the battery full and hot long-term, and swelling is only a matter of time. Better to take the battery out and wire power directly, or at least cap charging at 60%."

There's also a technical discount (inferred from the source, not tested): pairing signaling runs through public MQTT servers (free brokers from HiveMQ and EMQX), hole punching uses Google's STUN, and there's no relay server. So "no third-party servers" isn't entirely true, and behind symmetric NAT or some carriers' mobile networks the two phones may fail to connect at all.

### Who it's competing with

This category was already crowded ten years ago.

| Peers | Scale | Compared with alternatives |
|---|---|---|
| AlfredCamera | 65.6M installs on Google Play, Japan store "Home" Top Grossing #1 (11-06) | Requires an account, cloud recording, earns through subscriptions; the free version's picture quality is poor, exactly the complaint the author raised |
| IP Webcam | Google Play 28.4M installs | LAN streaming to a recorder for experts; going over the internet means DIY setup |
| WardenCam | Google Play 13.2M installs | Requires a Google account, recordings stored in cloud drive |
| Baby Monitor 3G, Bibino, Dormi | 440k–2.6M installs | Also two phones peer to peer, but built specifically as a baby monitor; paid and more polished |
| Haven(Guardian Project) | GitHub 6.8k stars, unmaintained since 2022 | A privacy-focused open-source veteran, emphasizing sensor intrusion alerts rather than live viewing |
| Frigate, secluso | GitHub 37k / 1.7k stars | Needs dedicated hardware such as a Raspberry Pi or a recorder |

CamMonitor adds nothing new in features and has far less than Alfred. What it holds is a small niche: an old-phone monitor that is **open source, free, needs no account, and is still maintained** . The niche is small, but it's enough for a good shareable pitch. So far I haven't seen anyone build another one along these lines.

### What to learn from it

- **Say what it saves you in one line** . "Old phone + no monthly fee + no cloud" fits in a single sentence that repost accounts can use as-is. In the same period, tv-detox (10-18 built, a tool that speeds up old Android TV boxes) picked up about 2200 stars, and linux-android, which runs Linux on an Android phone, got about 3800 stars. "Bringing old devices back to life" is a topic these accounts love to repost.

- **Have the download link ready** . Put ready-made APKs in Releases so the repost can link straight to the download page and readers can install in one tap.

- **The AI-coding build log is a seed story for the developer crowd** . The requirements draft, the spec, and every prompt were all public, which won the first stars and forks inside the author's own circle. But that story doesn't travel beyond the circle. What gets it out is the product line.

- **This kind of growth is one-off** . There's no demo video, no store listing, and no reason to keep updating, so once the reposts stop, the growth stops.

### If it were you

**Verdict** : it's not worth building another "old phone as a security camera" open-source project. The category is already crowded, and this one only got 135 stars. What's worth copying is the topic and the framing: find a combo that "everyone has lying around and everyone pays a monthly fee for", then build an open-source, free version.

**Who it suits** : developers who can finish a small tool in a week or two (using AI is fine), who care about topics like "old devices, subscription fees, privacy", and who are willing to polish the first line of the README and a promo image.

**Who it doesn't suit** : people who want a project that keeps gaining stars for the long term, since these projects stop growing after one wave of reposts. People who want an iPhone version: outside the EU, iOS doesn't allow direct installation of APK-style packages, and background camera access is heavily restricted, so repost accounts can't hand out a download link. And anyone who wants to take on Alfred; that's a different business.

**How to do it** : if you insist on the monitoring direction, fill the gaps it left: add a relay (TURN) to fix connection failures, add local motion detection and recording, add charge protection that stops at 60%, and put "stops your old phone from swelling up" in the README. An even better approach is to change the device or the subscription: an old tablet as a family photo frame or calendar display, an old phone as a baby monitor, an old TV box for ad blocking, an old laptop as a home NAS. Each one gets a line like "no more monthly fee for xx".

**How to get users** : first publish a long write-up of the process in your own circle (how AI built it, with the prompts attached) to get the first few dozen stars. Then send a one-line pitch, an image, and the APK or installer download link directly to X accounts that specialize in reposting GitHub projects, such as @RoundtableSpace and @GitHub_Daily. Send in both English and Chinese.

**Is now the window?** There's no rush. The "old devices brought back to life" topic has been reposted for years and doesn't depend on a specific time window. But the extra angle of "AI-written, I never read the code" is becoming more common, so telling it earlier still feels fresh (inference).

**Signs you've misread the market** : the pitch and image go out and two weeks pass without a single repost account picking them up. Or they do pick them up, reach tens of thousands of views, and the stars still don't pass 100. Either way it means the "saved monthly fee" isn't painful enough, and readers scroll past.

[GitHub repo ↗](https://github.com/samlu/CamMonitor)

**Research notes: competitor table, acquisition evidence, full research report**

Competitors

| Name | Where | Launched | Scale | How it differs from this one |
|---|---|---|---|---|
| AlfredCamera | iOS | About 2014 | Installed 65.6M times on Google Play, #1 on the Japan store's Home top grossing chart | Requires an account, cloud storage plus subscription, free version has poor image quality. CamMonitor is free and open source with no account needed, but has no motion-detection recording. |
| IP Webcam | Android | About 2011 | Play 28.4M installs | A LAN streaming service for power users to connect to an NVR. CamMonitor works by scanning a QR code and can go over the internet via P2P. |
| WardenCam | Android | About 2014 | Play 13.2M installs | Requires a Google account, recordings stored in cloud drive |
| Baby Monitor 3G / Bibino / Dormi | Android | Years | 440K–2.6M installs, paid | Same two-device P2P setup, but paid, more polished, and built specifically for the baby-monitor use case |
| Haven (Guardian Project) | Open source | 2017 | 6.8K GitHub stars, no updates since 2022 | Focuses on sensor-based intrusion alerts rather than live viewing, no longer maintained |
| Frigate / secluso | Open source | 2019 / 2024 | 37k / 1.7k stars | Needs an NVR or dedicated hardware such as a Raspberry Pi, so it's not an old-phone solution |
| Ring / Nest | Big-company features | — | — | Requires buying hardware plus a monthly fee, the exact comparison CamMonitor's copy is built around |

Where users come from

- **Verified** **Long Facebook post by the author (Taiwan dev circle)** : 10-19 "40-year veteran engineer's farewell: handed everything to AI", 40 comments; 10-19 forks, mostly from Taiwan accounts; stars reached 10-21, 26

- **Verified** **X English repost account @RoundtableSpace** : a 10-31 12:15 UTC post "no Ring subscription" linking to the Releases page, 77.6K views

- **Verified** **X Chinese recommendation account @GitHub_Daily** : a 11-01 14:30 UTC recommendation post, matching the 11-02/03 +31/+36 star spike

- **Verified** **Threads/FB reposts (0xalpha.ai, 一人公司研究所)** : traditional Chinese reposts around 11-02, same copy

- Inference **Paid promotion** : accounts like RoundtableSpace often charge for posts, but there's no evidence

## samlu/CamMonitor: why this wave of stars happened

### One line

CamMonitor is an Android open-source app built by Sam Lu, a veteran engineer in Hsinchu, Taiwan, with 40 years of experience, written entirely with AI (Antigravity CLI + Gemini). It pairs two Android phones by scanning a QR code: one acts as the camera, the other shows the live feed. No account, no cloud subscription, P2P encryption. Its success doesn't come from product strength. It's a **"old phone in a drawer + no Ring monthly fee + no cloud" story that's easy to tell** , reposted by an English repost account and a Chinese GitHub recommendation account on 10/31–11/1, which pushed stars from 30 to 135 within a week.

### Data

- Built 2026-10-18, Apache-2.0, Kotlin (about 11.7K lines). The author account samlu has only this one popular repo, with 10 followers (GitHub API, 2026-11-06).

- Star curve (GitHub repo_day): 10-21 26 → 10-24~10-31 flat at 30 → 11-01 35 → 11-02 66 → 11-03 102 → 11-04 124 → 11-05 134 → 11-06 135. gain_7d=105; only 7 in the first 4 days; single-day peak +36 (11-03); mostly stalled from 11-05. Never made trending.

- 24 forks in total. Early forks (around 10-19) were mostly Taiwan accounts (shengta, shunhsiung, cocoaswifty, bingo-taiwan, nelsonchung, etc.). Forks after 11-01 came from around the world (GitHub forks API).

- APK downloads (GitHub releases API, 2026-11-06): v1.0.6 154 times, v1.0.7 115 times, early versions around 30 combined. So the number of people who actually installed and used it is in the hundreds.

- Not on Google Play. The author stated "no plans to publish" and it's Android only.

### Where users come from (by time)

1. **10-19 作者自己的 Facebook 长文** (facebook.com/samlu128,「40 年老工程師的告別作:全交給了 AI」),40 条评论,是第一波 26 star 和台湾 fork 的来源。叙事重点是「我没看过一行代码,全是 AI 写的」,仓库里还放了需求草稿、规格书、全部提示词记录(UserPromptsHistoryUpTo1.0.2.md)和 AI 安全审计报告。

2. **10-31 12:15 UTC, X @RoundtableSpace (0xMarioNawfal)** : English post "You probably have an old Android phone sitting in a drawer. It can be a home security camera, with no Ring subscription…", linking directly to the Releases page, **77.6K views** (scraped from the x.com page via Jina). This was the largest exposure.

3. **10-31 14:36 UTC, X @qwerty__cs (拓跋鼠)** Chinese translation of the post, 488 views, little impact.

4. **11-01 14:30 UTC, X @GitHub_Daily (GitHubDaily)** Chinese recommendation "An old Android phone sitting unused in a drawer…". Replies asked for an iPhone version, and someone warned that 24 hours of streaming would make the battery swell. The timing matches the star peaks of 11-02/11-03 and the two days of +31/+36 stars very closely.

5. **11-02, Facebook page "一人公司研究所"** traditional Chinese repost (10 comments). **Threads @0xalpha.ai** "CamMonitor turns an old phone into a monitor, no Ring monthly fee". All use the same copy (no account / no cloud / 720p / scan a QR code), clearly translated and reposted from one another.

6. The author 11-01 added a Japanese README and 11-02 added a Japanese interface, but I found no evidence of spread within the Japanese community.

7. Not found: HN, Reddit, Product Hunt, press coverage, TikTok. Searching CamMonitor in the local x-radar corpus returned 0 results.

Conclusion: the growth came from **one-off reposts by content-repost accounts** , not sustained organic traffic. When the reposts stopped, the stars stopped too.

### Competitors

| Competitors | Where | Launched | Scale | How CamMonitor differs |
|---|---|---|---|---|
| AlfredCamera | Android/iOS | Around 2014 | 65.6M installs on Google Play, +100K in 7 days; #1 on the Japan store's Home top grossing chart (googleplay library) | Requires an account, has cloud storage, free version has poor image quality and ads, makes money from subscriptions; full feature set (motion detection, recording, cloud storage) |
| IP Webcam | Android | Around 2011 | Play 28.4M installs | LAN MJPEG/RTSP service for power users to connect to an NVR; going over the internet requires DIY setup |
| WardenCam | Android/iOS | Around 2014 | Play 13.2M installs | Requires a Google account, recordings stored in Dropbox/Drive |
| ZoomOn / Security Camera CZ / SeeCiTV / Camerito | Android | Years | 290k–5M installs | Same "one phone watches, one phone records" model, but most have accounts, ads, or subscriptions |
| Baby Monitor 3G / Bibino / Dormi | Android/iOS | Years | 440K–2.6M installs, paid and profitable | Baby-monitor use case, same two-device P2P approach, but paid and more polished |
| Haven (Guardian Project, endorsed by Snowden) | Open source | 2017 | 6.8K GitHub stars, unmaintained since 2022 | Open-source privacy pioneer; focuses on sensor-based intrusion detection and Signal alerts rather than live viewing |
| Frigate / clearcam / secluso | Open source | 2019 / 2025 / 2024 | 37k / 2.2k / 1.7k stars | Requires an NVR or dedicated hardware like a Raspberry Pi, not an "old phone" solution |
| Ring / Google Nest, etc. | Big-brand hardware | — | — | Requires buying hardware plus a monthly fee, the exact thing CamMonitor's copy uses as its comparison |

CamMonitor has nothing new in features. The "old phone as a camera" category was already crowded a decade ago, and Alfred alone has 65M installs. The author admits in the post that it's "an old topic, with plenty of existing apps".

### Differentiation

- **No account + open source + free** : most comparable store apps require login, carry ads, or charge subscriptions; open-source comparables either stopped updating (Haven) or need dedicated hardware. "No sign-up, no cloud, scan once" is exactly the core line of the repost copy.

- **Technical implementation** : WebRTC P2P direct connection, signaling over public MQTT (broker.hivemq.com / broker.emqx.io, TLS + ECDSA signatures), hole punching via Google STUN, **no TURN relay** . So "no third-party server" isn't strictly true, and it may fail to connect behind symmetric NAT or some mobile networks (not tested, inferred from the source code).

- **Recording and motion detection are deliberately left out** (to prevent overheating and battery swelling). It only offers "live view on demand + two-way audio + remote flashlight", with far fewer features than Alfred.

- **The story is the difference** : "A 40-year veteran engineer's farewell + never read a line of code + all prompts published". This was the real selling point of the first wave on Facebook. It attracted developers who want to see an AI coding case study, not people who want surveillance.

Is differentiation why it took off? **Partly** : repost accounts picked it because "no Ring monthly fee, no cloud, old phone" are easy lines to write and spread, and open source with no account gave them exactly those lines. But the stars came from the reposting accounts' distribution, not from users spreading it on their own (APK downloads were only a few hundred, and stars stopped climbing once the reposts stopped).

### Visibility

It's been circulating in small circles: an English repost thread with 77.6K views and a GitHubDaily shout-out, but no media, HN, or Reddit. The repo has 135 stars and no copycats (searching GitHub for the same name only turns up unrelated older projects, and the 24 forks have no derivative work with stars). No revenue involved.

### What to borrow

- "Revive old devices + skip subscription fees + no cloud" is a topic X repost accounts love (around the same time, Ansub/tv-detox, "making old Android TVs fast with Claude Code," hit 10-18 build 2192 stars, and mayukh4/linux-android got 3845 stars). Build a tool like this, put "no account / no monthly fee" in the first line of the README, add a promo image, and accounts like @RoundtableSpace and @GitHub_Daily will pick it up easily.

- Adding "written by AI, prompts fully public" as a side story can win the first batch of seed stars in developer circles.

- Ship ready-made APKs in Releases so repost threads can link straight to the download page.

- This kind of burst is one-off. To sustain it, you need to get into the app stores or solve real needs like NAT traversal and recording, which the author has explicitly said they won't do.

### Uncertain

- The timing of the star daily snapshots is unclear, so we can't tell whether the peak of 11-02/03 came mainly from RoundtableSpace or GitHubDaily. Both are possible.

- Stargazer timestamps and regional breakdowns weren't available (API 403/404).

- Whether RoundtableSpace's reposts were paid promotion is unknown (these accounts often charge, but there's no evidence).

- Actual connection success rates and retention are unknown.

### Sources

- GitHub API: repos/samlu/CamMonitor, releases, forks, commits; README, WhyCreateThisProject.md; source grep (MQTT broker, STUN)

- Our github library repo_day / repo_profile; googleplay library app_profile (2026-11-06)

- [x.com/RoundtableSpace/status/2116504206614114784](https://x.com/RoundtableSpace/status/2116504206614114784)

- [x.com/qwerty__cs/status/2116539913596453113](https://x.com/qwerty__cs/status/2116539913596453113)

- [x.com/GitHub_Daily/status/2116900590408355924](https://x.com/GitHub_Daily/status/2116900590408355924)

- [www.facebook.com/samlu128/posts/10227309513482499/](https://www.facebook.com/samlu128/posts/10227309513482499/)

- [www.facebook.com/61588107756334/posts/122125472169270258/](https://www.facebook.com/61588107756334/posts/122125472169270258/) (One-Person Company Research Institute)

- [www.threads.com/@0xalpha.ai/post/DfK-EXyE8uN](https://www.threads.com/@0xalpha.ai/post/DfK-EXyE8uN)

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