# Scan an office with an iPhone, then have Codex rebuild it from the photos into a walkable web page: Wes Bos's 2800-star repo, the stars came from the video · GitHub open-source project

> Source: https://stringers.io/en/github/gh_wesbos__3d-model-anything.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

> Well-known JS instructor Wes Bos posted an office video on X, YouTube, TikTok, and his blog on the same day, where you walk through the office and shoot things with a light gun. The X main post hit 500k views; then a dozen-plus AI news sites and Chinese recommendation accounts reshared it.

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

wesbos/3d-model-anything · wesbos

# Scan an office with an iPhone, then have Codex rebuild it from the photos into a walkable web page: Wes Bos's 2800-star repo, the stars came from the video

**What it is** An open-source workflow: scan a room with an iPhone that has LiDAR, take photos with camera positions, then have a coding agent rebuild it round by round, guided by the photos, into an editable Three.js scene you can walk through in the browser. It includes a capture app and an online demo.

**Why it took off** Well-known JS instructor Wes Bos posted a video of walking through an office and shooting things with a light gun on X, YouTube, TikTok and his blog on the same day; the X main post got 500k views. Over a dozen AI news sites and Chinese recommendation accounts reshared it afterward.

**C** **How you could follow** Stop doing room-scan-to-Three.js; take the loop of "measurement + photos with camera positions + the agent checks its own screenshots against them and revises" and apply it to turning single objects into CAD, Quest spatial scans into WebXR, or adding "help me rearrange this" to a rebuilt room. Shoot a before/after comparison video first, then pitch it to recommendation accounts.

Grade C: the method is worth learning, but the stars mostly came from the author's own audience, which followers can't reproduce. A similar-idea project that launched earlier, 3d-home-tv-os, has only 58 stars, and the same topic has already been widely seen in developer circles. Already public / widely discussed

Data (updated daily) · Repo created 2026-10-21 · 2,839 stars · 7 days +307

GitHub · stars

Data 2026-10-22 → 2026-11-08 (18 days, 18 observations). Points before 10-28 come from deeply / archive backfill, not every day has one

In Oct 2026 22, JavaScript instructor Wes Bos posted a video titled "my office is complete." He scanned his office with an iPhone, then had Codex rebuild it as a Three.js scene in the browser: desk, bookshelves, a 3D printer, a whole wall of filament spools, all walkable, and hooked up to a physics engine so he can knock things flying with a Nintendo light gun. The repo wesbos/3d-model-anything is the "take it and build your own" link under the video; in 16 days it gained 2792 stars and 413 forks. This piece covers why it spread, what's new about the method, and what you can learn, and what you can't, if you want a project to get this many stars too.

### One video, four platforms, same day

The main post on X went up in the afternoon of 10-22, and by 11-06 it had 508k views, 4314 likes and 6600 bookmarks. Bookmarks outnumber likes, and many people's reaction was "saving this to try on my own room later." The same day, his YouTube video (11-06k views as of 127), TikTok (47k views) and blog post went out together. The X main post links to YouTube, not the repo; the repo is linked in the video description and the blog.

The blog post opens with these two sentences:

> I scanned my office with my iPhone, then used Codex to rebuild it in Three.js. ... Then, naturally, I added physics so I could throw everything around. And a Nintendo Zapper.

Wes Bos needs no introduction in developer circles: host of the Syntax podcast, many years of selling JavaScript courses, 447k followers on X, and 79 repos on GitHub with 84k stars combined. That context explains all the numbers that follow.

### What it actually is

The repo contains four things: an iPhone capture app (source code; it's also free on the App Store starting at 10-24), a Three.js browser viewer, an agent skill called `scan-to-threejs`, plus a complete example of his office.

The key is what the phone scan actually produces. The iPhone scan doesn't give you a finished model; in his words it's "evidence":

- RoomPlan gives rough boxes for walls, doors, windows and furniture, handling the layout;

- LiDAR gives surface meshes, handling dimensions;

- photos handle appearance: materials, labels, cables, what's on the shelves;

- each photo records the phone's position, orientation and lens parameters at the moment it was taken.

The last point is the core of the project. Because each photo's camera position is known, the agent can place the camera in Three.js at the same spot, render a frame, and compare it side by side with the original photo: is the desk too wide, is the shelf off, does the chair look right. Then it edits the code, takes a screenshot, and compares again. Wes himself says this isn't a one-click conversion: "getting the layout right is one round; materials, lighting and small details are a whole other job."

The output is editable Three.js code, and every object is a separate piece you can pull out on its own. So you can switch to wireframe, go top-down, lay every object out flat (knolling), or drop them into a physics engine.

### How the stars grew

| Date | Cumulative stars | What happened that day |
|---|---|---|
| 10-21 | 0 | Repo created |
| 10-22 | 62 | Same day launch on X, YouTube, TikTok and the blog |
| 10-23 | 152 | First AI news sites start writing about it |
| 10-24 | 968 | +816 in one day, the biggest day; App Store version goes live |
| 10-25 | 1507 | +539, a Taiwanese tech site and a Chinese-language X account start sharing it |
| 10-26 | 1866 | +359, picked up by daily.dev |
| 10-27 | 2036 | +170, Wes posts that the App is live (68k views) |
| 10-28 to 11-06 | 2244 → 2792 | 50 a day or so; 11-01 GitHub_Daily posts a Chinese recommendation (92k views) |

(Daily GitHub snapshot, Beijing time; 11-06.)

The peak hit the day after launch, then roughly halved each day: 816, 539, 359, 170. That's the classic shape of a one-off content spike, not reputation built up slowly. The second wave came from a dozen-plus AI news sites that automatically pick up fast-rising new repos (Reveneau, EarlySignal, AlphaSignal, daily.dev, Trendshift and others), followed by the Chinese-language crowd: Sina, Zhihu, Juejin's hands-on write-up, CSDN, and GitHub_Daily on X. I found no posts about it on Hacker News or Reddit.

The fork-to-star ratio is about 15%, higher than comparable new repos (median around 11%), which fits a usage pattern of "clone it and let an agent run it." But few people have actually scanned a room with it: the App Store has only 11-06 ratings as of 2. Many stars came from "watched the online demo," not "got it running."

### A new method, but the stars don't come from the method

Looked at individually, every piece has been done before:

| Who | What it is | Scale | Compared with alternatives |
|---|---|---|---|
| Apple RoomPlan and apps built on it | A system-level room-scanning API since 2022 | Works on every LiDAR iPhone | Outputs only boxes; treats RoomPlan as one of its inputs |
| Polycam, Scaniverse, Luma | 3D scanning apps | Mature product | Output meshes or Gaussian splats; they look right but can't be edited |
| brunosimon/my-room-in-3d | "My room" hand-modeled in Blender | 4532 stars (2021). | Same selling point, all done by hand |
| viettranx/3dviz-pro-max and other Three.js skills | Generate Three.js scenes from text or images | 1040 stars | No measurement, no camera-pose matching, so they can't reproduce a real room |
| reality-opened/openreality | Phone video turned into a 3D scene that agents can query | 620 stars | Aims to help agents understand space, not to rebuild it beautifully |
| Rod292/3d-home-tv-os | Home scan → 3D → TV main screen | 58 stars | Also uses RoomPlan + LiDAR, and started earlier |

(Star counts are from the 11-06 GitHub snapshot.)

The last row matters most. 3d-home-tv-os, which is close in approach and was posted earlier, is stuck at 58 stars. The chain of "LiDAR for measurements + photos with camera poses + the agent checking its own renders against them" was publicly put together first by Wes, and he was the first to get lots of eyes on it. But the scale came from his name and that polished video. The method decides how many people in this wave bothered to bookmark or fork it. The Juejin hands-on write-up specifically praised it for "writing its own error into the JSON."

I haven't seen a repo that copies it directly. Someone used it to build a live digital twin of their own office. Commenters are asking for an Android version and a Quest version, and some want "help me rearrange the room."

### What to learn from it

- **Content first, repo second** . A flashy video + an online demo you can open right away + an open-source repo + a free App, all launched the same day. The stars were a byproduct of the content. Without your own audience, you'd need to swap this for getting big accounts to share it.

- **Feed the model structured evidence, then let it check its own work against that evidence** . One reason "LLMs can't do 3D" is that the model can't see where it went wrong. Give it measurements and same-pose photos, and it has something to correct against. The idea transfers: scan an object and output CAD, scan a machine and output an assembly drawing, scan a house and output a renovation plan.

- **Ship the skill bundled with your capture tool** . Anyone can copy a single skill file. A full workflow only comes together when you include your own capture App and data format.

### If it were you

**Verdict** : worth learning the method, not worth going head-to-head. Building another "room scan → Three.js" won't get this kind of star count. 3d-home-tv-os sitting at 58 stars shows what happens without traffic, and developer circles have already looked at this idea once.

**Who it's for** : developers who already have some name recognition in a 3D or hardware niche and can cut short videos; people who own devices with depth sensors like a Quest or an iPad Pro.

**Who it's not for** : anyone without any distribution channel who expects the project to earn stars on its own; anyone hoping to repeat the same topic (rooms, Three.js), since this has already taken that spot.

**How to do it** : keep the loop of "measure + photos with camera poses + agent compares and iterates," but change the input or the output. Three directions: a single object → editable CAD or Blender model (not a mesh); Quest 3 spatial scans → WebXR scenes (people are asking for this); add "help me rearrange it" on top of a rebuilt room so the agent proposes several layouts. Put a GIF in the first screen of the README showing the original photo and the render side by side, so people immediately see that it finds the differences on its own.

**How to get users** : make the video first, then write the README. Shoot your own real objects or room, with before-and-after comparisons and something interactive (throw it, take it apart, change colors). When posting on X, link the video, not the repo. At the same time, send the video to Chinese-language recommendation accounts like GitHub_Daily and English AI news sites. Those were the source of this second wave of traffic.

**When before** : within a few months (my estimate). "Agents + skills" is a hot term right now, and the RoomPlan and LiDAR barriers still keep a lot of people out. Once scanning apps build in "export as an editable scene," this demo won't stand out anymore.

**Signs this read is wrong** : the video hasn't reached tens of thousands of views a week after it went out, and no recommendation account picks it up; or Polycam or Apple already output editable structured models directly; or a repo in the same niche on GitHub already has several hundred stars.

[GitHub repo ↗](https://github.com/wesbos/3d-model-anything)

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

Competitors

| Name | Where | Launched | Scale | How it differs from this one |
|---|---|---|---|---|
| Apple RoomPlan and Magicplan / Canvas, etc. | Big-company features | 2022(iOS 16) | A system API, works on every LiDAR iPhone | RoomPlan only gives walls, doors, windows and furniture boxes; this tool treats RoomPlan as one input and has an agent use photos to fill in the real objects |
| Polycam | iOS | 2020 | Leading 3D scanning apps (unverified) | Output scan meshes / splats that can't be edited by object; this outputs code objects you can pick up individually, with physics |
| Scaniverse / Luma AI | iOS | 2021 / 2022 | Free, large user base (unverified) | Gaussian splatting / NeRF looks realistic but isn't structured geometry and can't be edited |
| Matterport | Other | 2011 | A leader in real-estate digital twins, paid | Professional hardware + SaaS aimed at real estate; this is a free, developer-facing DIY workflow |
| reality-opened/openreality | Open source | 2026-09-01 | 620 stars | Phone video → 3D scene that agents can query + MCP, focused on spatial understanding, doesn't rebuild an editable scene |
| Rod292/3d-home-tv-os | Open source | 2026-10, mid | 58 stars | Also scans homes with RoomPlan + LiDAR for 3D, but with no author traffic, stuck at 58 stars |
| brunosimon/my-room-in-3d | Open source | 2021 | 4532 stars | A room demo hand-modeled in Blender; this one is automatic reconstruction from a scan with an agent |
| viettranx/3dviz-pro-max and other Three.js agent skills | Open source | 2026-09 | 1040 stars | Generates scenes from text or images, with no real measurements or camera-pose matching, so they can't reproduce a real room |

Where users come from

- **Verified** **First posted by X author** : 10-22 main post with 508K views, 4314 likes, 6600 bookmarks; the author has 446,846 followers; 10-27 App listing post with 68K views

- **Verified** **Author posts on YouTube / TikTok / blog** : YouTube "my office is complete" 127k views; TikTok 47k views; wesbos.com blog 10-22

- **Verified** **AI news aggregators / fast-rising star lists** : Reveneau (10-23), EarlySignal ("goes viral" 10-24), LockerGnome, AlphaSignal, daily.dev, Trendshift daily chart #14, and others

- **Verified** **Chinese-language reposts** : New Mobile Life, Threads, Sina, Zhihu, Juejin, CSDN; on X, GitHub_Daily post 11-01 with 92K views

- **Verified** **Free on the App Store** : 10-24 listed on 1.0, free, with only 2 ratings so far

- Inferred **Hacker News / Reddit** : HN Algolia and Reddit both returned no relevant posts

## Why wesbos/3d-model-anything runs as it does

> Channel: GitHub. Research date: 2026-11-06. Sources: our GitHub database (repo_profile / repo_day), data cards, web pages (Jina / Exa search), the fxtwitter public endpoint (X post data), TikHub (TikTok), and iTunes lookup.

### Bottom line

This is an "open-source repo attached to creator content": Wes Bos (447k X followers, Syntax podcast host, well-known JS instructor) first posted a slick video, "my office is complete," in which he scanned his own office and had Codex rebuild it as a Three.js scene you can walk through and throw things in. The repo is the video's "take it and build your own" entry point. The stars came mostly from his own distribution, then got passed around by dozens of AI news aggregators and Chinese GitHub recommendation accounts. The product's novelty is the workflow: **LiDAR plus camera-posed photos plus an agent skill, so a coding agent screenshots from the original camera positions, compares against the photos, and iteratively fixes the model** . The method is genuinely new, but the speed of the stars mainly came from the author's own reach.

### 1. Data shape

| Date | star | Daily gain | Event |
|---|---|---|---|
| 10-21 | 0 | — | Repo created (created_at 2026-10-21 19:38 UTC) |
| 10-22 | 62 | — | Blog, YouTube, the main X post, and TikTok published the same day (X post at 12:43 UTC) |
| 10-23 | 152 | +90 | AI news sites like Reveneau started writing; Reveneau says the GitHub API showed 667 stars that day and Trendshift's daily chart #14 (our database uses a Beijing-time snapshot, so the counting basis differs from theirs) |
| 10-24 | 968 | **+816** | Peak single day; App Store version listed 1.0 that day; EarlySignal "first spotted" it at 1.4k stars |
| 10-25 | 1507 | +539 | Taiwan's "New Mobile Life" and Chinese X accounts started sharing it |
| 10-26 | 1866 | +359 | Indexed by daily.dev |
| 10-27 | 2036 | +170 | Wes posted the App Store launch announcement (68k views) |
| 10-28→11-06 | 2244→2792 | about 50 per day on average | Long tail: 11-01 GitHub_Daily (92k followers) posted a Chinese recommendation |

Source: data card / github database repo_day. Currently 2792 stars and 413 forks, a fork/star ratio of about 14.8% (the median for new repos with ≥5000 stars is about 11%, so this is high, consistent with people cloning it to run with an agent). Last 7 days +435, prior 7 days +2205; the repo's status is "cooling." No suspicious flags (suspect is empty).

### 2. Competitors: who else does the same thing

"Scan a room with your phone and get a 3D model" is itself a well-worn idea. What's different is the **output** : others give you a mesh, point cloud, Gaussian splat, or floor plan, while this gives you **editable Three.js code** , written by an agent working from the photos.

| Product | Where | Launched | Scale | How it differs |
|---|---|---|---|---|
| Apple RoomPlan (API) and products built on it such as Magicplan and Canvas | Big-company features / iOS | RoomPlan 2022(iOS 16) | A system-level API that works on every LiDAR iPhone | RoomPlan only gives box shapes for walls, doors, windows, and furniture; 3d-model-anything uses RoomPlan as one input, then has an agent use photos to fill in real object details |
| Polycam | iOS / Android / web | 2020 | Leading 3D scanning apps (scale not verified) | They output scan meshes or splats, a "photorealistic shell" that can't be edited object by object; this outputs code objects you can pick up individually, with physics |
| Scaniverse (Niantic), Luma AI | iOS / web | 2021 / 2022 | Free, large user base (unverified) | Gaussian splatting / NeRF, which look realistic but aren't structured geometry and can't be edited |
| Matterport | Professional SaaS | 2011 | A leader in real-estate digital twins | Professional cameras plus payment, aimed at real estate; this is free and DIY for developers |
| reality-opened/openreality | Open source | 2026-09-01 | 620 stars (github library) | Phone video to an "AI-queryable" 3D scene plus MCP tools, focused on helping agents understand space rather than rebuilding a polished, editable scene |
| 0xProbabillity/house-control | Open source | 2026-10-14 | 872 stars | Turning your home into a 3D model to control smart-home devices; the focus is control, not reconstruction |
| Rod292/3d-home-tv-os | Open source | Mid-2026-10 (earlier than this one) | 58 stars | Scan home → 3D → home TV main screen, also using RoomPlan + LiDAR, but without the author's reach; stuck at 58 stars |
| brunosimon/my-room-in-3d | Open source | 2021 | 4532 stars | A classic demo of hand-modeling "my room in 3D" in Blender; it touches the same idea of moving your own room into the browser, but it's all manual |
| viettranx/3dviz-pro-max, img2threejs/plugin-interior and other Three.js agent skills | Open source | 2026-09/10 | 1040 / 258 stars | Generate Three.js scenes from text or images, with no real measurement or camera-position matching, so they can't recreate a real room |

Bottom line: taken separately, every piece has been built by someone (scanning apps, RoomPlan, Three.js generation skills). But the chain of "LiDAR measurement + posed photos + an agent that renders from the original camera positions, compares to the photos, and iterates" is the first version of this stack, open source, to get wide attention. The similar Rod292/3d-home-tv-os, from the same period, has only 58 stars, suggesting reach is the dividing line.

### 3. Growth tactics: where users come from

1. **Author's own multi-platform launch (verified, the main driver)** : 2026-10-22 all published on the same day.
    - Main X post (10-22 12:43 UTC): **508k views, 4314 likes, 232 reposts, 6600 bookmarks** ; the author has 446,846 followers (fxtwitter endpoint). The main post links to the YouTube video, not the repo; the repo is in the video and blog.
    - YouTube "my office is complete": **127k views** (the page shows 2 weeks ago, and is labeled Includes paid promotion, meaning the video had a commercial sponsorship, not that the repo was bought traffic).
    - TikTok @wesbos video with the same title: 47k views, 1647 likes (TikHub); also a YouTube Short.
    - Blog post wesbos.com/3d-model-anything (10-22).
    - 10-27 A second post announced the App Store launch: 68k views, 1311 likes.

2. **Auto-reposted by AI news and aggregator sites (verified)** : Reveneau (10-23, AI-written), EarlySignal (10-24, "goes viral"), LockerGnome, Fervor AI, AlphaSignal, daily.dev, Trendshift, rising.dev, BestHub, AC0.AI, Norvik Tech and more than a dozen others. Most automatically pick up "new repos gaining stars fast," then write articles with AI, and the star growth itself brings more exposure.

3. **Secondary spread in the Chinese-language sphere (verified, moderate volume)** : Taiwan's New Mobile Life (10-24), Threads, Sina, Zhihu ("just-open-sourced projects whose stars are rising fast, 6 of them"), Juejin hands-on articles, CSDN reviews; on X, GitHub_Daily (11-192k views, 4146 likes). Part of the long tail after 10-28 probably came from here.

4. **Hacker News / Reddit: none found** . HN Algolia returns no posts about this repo; Reddit search also came up empty. Our x-radar local corpus has zero hits for these terms (coverage is limited, so this doesn't mean nobody discussed it on X).

5. **GitHub trending: not listed** (repos under 60 days old can't make trending, so this doesn't indicate anything).

### 4. Why it took off

- **The author's reach was the primary driver** : Wes Bos has 79 repos in our database, 84k stars total, and 447k X followers. Four platforms (X, YouTube, TikTok, blog) launched together on the same day. The biggest single-day gain (+816) came the day after launch, and each day after that roughly halved (539→359→170). That's the typical curve of a one-time content spike, not organic word-of-mouth growing steadily.

- **The demo itself travels well** : the scanned office can be walked through, switched to wireframe, top-down, or knolling view (laying objects out flat), and even run through a physics engine to shoot cables with a Nintendo Zapper. 6600 more people bookmarked it than liked it (4314 likes), meaning many saved it to try later. That's exactly where the stars came from.

- **Timing: coding agents and skills were a hot topic** : in autumn 2026, the Codex and Claude Code skill ecosystems were booming, and many "Three.js + agent skill" repos were gaining stars at the same time (see the competitor table). Its pitch, "LLMs have never been good at 3D; feed them measurements and camera-posed photos," (Wes's words: "I have discovered the BEST way to get 3D modeling results from LLMs") answered a question developers were asking at the time.

- **Low enough barrier to watch** : the online demo (wesbos.github.io/3d-model-anything) works as soon as it opens; the app is free on the App Store 10-24; phones without LiDAR can still import and view models. Many stars came from "watched the demo" rather than "got it running." The App Store has only 2 ratings (iTunes lookup, 11-06), so very few people have actually scanned their own room.

### 5. What's the differentiation, and is it the reason it took off?

The differentiation is in the **method** : treat the phone scan as "evidence" rather than a finished product. LiDAR handles dimensions; photos with camera poses and intrinsics handle appearance. The agent writes Three.js, screenshots from the original camera position, compares side by side with the original photo, and revises, forming a self-checking loop. The output is editable, semantic code objects, not a mesh. This differs from Polycam, Scaniverse, and RoomPlan, and it's the point that keeps coming up in X bookmarks and Chinese technical articles (Juejin: "it writes its own error into the JSON").

But it's **not the main reason it took off** . Rod3d-home-tv-os, which also uses RoomPlan + LiDAR for home 292/3D, has only 58 stars. Wes Bos's distribution and one beautiful video set the scale; the differentiation determines how many people in that wave bookmarked or forked it (the fork ratio is high).

### 6. Visibility and copycats

- **How widely it's been seen** : it has been widely discussed in developer circles. The main X post has 500k views, YouTube 127k, and more than a dozen English AI news sites and Chinese outlets (Sina, Zhihu, Juejin, CSDN, GitHub_Daily) have covered it. No revenue involved (open source plus a free app).

- **Copycats** : no direct clones found yet (searching GitHub for scan-to-threejs / 3d model anything / roomplan turns up only this repo from the past 30 days; the other is Rod292, which is even older). Users have already built their own versions, such as @moeealii's "The Office — a live digital twin" (office.thealifamily.ca), and people are asking for Android and Quest versions.

### 7. What to borrow

- **"Content first, repo last"** : a flashy video plus an online demo plus an open-source repo plus a free app, all posted the same day across platforms. The stars were a byproduct of the content. Without the author's reach, this playbook would need to rely on getting big accounts to repost.

- **"Feed LLMs structured evidence and have them check their own work"** : this idea can carry over to other fields (scanning objects into CAD, scanning workshops into digital twins, scanning houses into renovation plans). Wes's reply section already has people asking for features like "organize my room / redesign my room," a product direction that could be built on.

- **Bundle the skill with a companion capture tool** : a skill alone is easy to copy. Pairing it with your own capture app and data format (house-capture/5) creates a complete workflow.

### 8. Still unclear

- How many people have actually made it work on their own room: the App Store has only 2 ratings, and there's one hands-on article on Juejin, so the scale can't be estimated.

- What triggered the 10-23/10-24 spike: our snapshot uses Beijing time, which doesn't match Reveneau's "10-23 667 stars" count, possibly a counting-basis difference. Whether other big accounts reposted it on 10-24 besides the App Store launch hasn't been checked.

- The contents of wy51ai/floorplan-3d (3582 stars, created 2026-09-29, no description, README unavailable) haven't been verified; it may be a similar project.

- The current scale of commercial competitors such as Polycam and Scaniverse hasn't been verified.

### Sources

- README on GitHub: [github.com/wesbos/3d-model-anything](https://github.com/wesbos/3d-model-anything)

- Blog: [wesbos.com/3d-model-anything](https://wesbos.com/3d-model-anything) (2026-10-22)

- YouTube: [www.youtube.com/watch?v=JGQCcCzuS88](https://www.youtube.com/watch?v=JGQCcCzuS88) (127k views, 11-06 fetched)

- X main post: [x.com/wesbos/status/2113249938125553867](https://x.com/wesbos/status/2113249938125553867); App launch post: [x.com/wesbos/status/2115097903169683605](https://x.com/wesbos/status/2115097903169683605) (data from api.fxtwitter.com, 11-06)

- TikTok:[www.tiktok.com/@wesbos/video/7699217763599944978](https://www.tiktok.com/@wesbos/video/7699217763599944978)(TikHub,11-06)

- App Store: id6825843398,2026-10-24 listed, 2 ratings (iTunes lookup, 11-06)

- Reveneau:[reveneau.com/ainews/wesbos-3d-model-anything-iphone-lidar-th…](https://reveneau.com/ainews/wesbos-3d-model-anything-iphone-lidar-three-js-666-stars-two-days)

- EarlySignal:[earlysignal.news/s/wesbos-3d-model-anything](https://earlysignal.news/s/wesbos-3d-model-anything)

- New Mobile Life:[www.newmobilelife.com/2026/10/24/3d-model-anything-iphone-li…](https://www.newmobilelife.com/2026/10/24/3d-model-anything-iphone-lidar-threejs/)

- Zhihu: [zhuanlan.zhihu.com/p/2099562531752878207](https://zhuanlan.zhihu.com/p/2099562531752878207); Juejin: [juejin.cn/post/7700894865731715135](https://juejin.cn/post/7700894865731715135)

- GitHub_Daily post: [x.com/GitHub_Daily/status/2116764687920746539](https://x.com/GitHub_Daily/status/2116764687920746539) (92k views)

- Competitor star counts: our GitHub repo_profile (2026-11-06)

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