研究内容为英文. 本页的研究内容（标题、简报、来源）以英文撰写，未经机器翻译。只有界面已翻译。

新想法Consumer and prosumer 3D

3D-Model-Anything-powered room-to-3D tour web…

2周前 · 评分于 2026年10月24日 · ID HBV-261024-05

暂定标题： 3D-Model-Anything-powered room-to-3D tour web app (working name) — upload an iPhone room scan, get an editable, shareable 3D room you can walk through in the browser

摘要

简报假设估算

Benchmark: wesbos/3d-model-anything — total 1,338 stars (gh api / OSS Insight, fetched 2026-10-24 07:02 BRT; browser-confirmed (abbreviated 1.3k, repo page 07:19–07:30 BRT), exact from API/OSS Insight); created 2026-10-21 (age 3 days); 24h (23 Oct UTC) +578; 7d (17–23 Oct) +1,046 vs prior 7d (10–16 Oct) +0 = n/a (n/a (…

阅读其余部分

no prior)); 3d (21–23 Oct) +1,046 vs prior 3d (18–20 Oct) +0 = n/a (n/a (no prior)); since creation 446.0/day; peak day 2026-10-23 (+578), 0 days since peak; license NOASSERTION; forks 220. Daily stars (OSS Insight): 21 Oct +1, 22 Oct +467, 23 Oct +578, 24 Oct partial +267. Description: "Capture iPhone LiDAR and photos, reconstruct editable Three.js rooms, and explore them in a browser." License: code and skill MIT; office demo images/textures under separate ASSETS-LICENSE (GitHub shows NOASSERTION because of that split) — commercial hosting of the code is allowed with the MIT notice. Method: daily new stars from OSS Insight analyze-recent-stars (last 28 days + prior 28, UTC days) and totals from the GitHub API (gh api repos/OWNER/REPO, cross-check), fetched 24 Oct 2026 ~06:59–07:40 BRT; totals appended to /workspace/hbv_star_snapshots.csv with source. 24h = 23 Oct UTC; 7d = 17–23 Oct vs 10–16 Oct; 3d = 21–23 Oct vs 18–20 Oct; label: >+10% accelerating, <−10% decelerating, otherwise steady; peak over the ~56 days available. Browser check (logged-out github.com, 24 Oct 2026): trending pages read 06:58 BRT (weekly all languages, daily, Python weekly, TypeScript weekly) list none of these repos; repo pages read 07:19–07:30 BRT confirm the totals (browser-confirmed (abbreviated), exact from API/OSS Insight); star-history charts were read for shape only, not values. Star-history shape (browser, approximate axis read, shape only): near-linear steady rise from ~zero on Thu 22 Oct to the current total on Sat 24 Oct; oldest commit 21 Oct 2026; sidebar shows no named license (README: source and skill MIT, separate ASSETS-LICENSE.md). None are estimated. Existence check: Checked 24 Oct 2026: the maintainer (Wes Bos) ships the iPhone app as source only ("an App Store release is a separate distribution step") and does not sell a hosted version; no hosted wrapper of this repo was found in this run's GitHub/web checks (wrapper count: none found; forks 220 are not products). Hosted room scanning already exists from Polycam, Matterport, magicplan, Luma and RoomPlan-based apps. X: 35 posts/7d (Up from zero 0% Sun→Fri (full UTC days: Sun 18 0 → Fri 23 19); peak Fri 23 19; Fri 23 is 0% below peak; window 2026-10-17 10:02–10:04 UTC (start varies by query) → 2026-10-24 ~10:03 UTC (X API 7-day limit; Sat 17 and Sat 24 are partial UTC days; full days Sun 18 → Fri 23)). Revenue hypothesis: ~$500–$3K/month at month 6 (HYPOTHESIS: 100–300 rooms/month × ~US$12).

描述假设

A hosted, no-terminal version of Wes Bos's 3D Model Anything: scan a room on a LiDAR iPhone, upload the export, and an AI agent rebuilds it as an editable Three.js scene with dollhouse and walk-through views and a share link. Who: Home renovators, Airbnb hosts, small real-estate agents and interior-design hobbyists with LiDAR iPhones who want a shareable 3D room without learning 3D tools.…

阅读其余部分

Pain: 3D Model Anything needs Xcode to install the capture app, Node and Python locally, and an AI coding agent driving an iterative skill; non-engineers cannot use it. Monetization: Pay per room (hypothesis ~US$9–19) or Pro around US$15/month for unlimited rooms and branded share links; MIT allows commercial hosting with attribution. Build: 3–4 agent-weeks (upload intake for house-capture exports, sandboxed agent runs of the scan-to-threejs skill, viewer hosting, share links, Stripe credits); stack: Next.js on Vercel; sandboxed Claude Agent SDK runner; Three.js viewer from the repo (MIT); Vercel Blob/R2 for scans; Stripe. GTM: Shareable 3D room links (every tour is an ad); real-estate and Airbnb host communities; SEO for "scan room to 3D model free".

目标客户

Home renovators, Airbnb hosts, small real-estate agents and interior-design hobbyists with LiDAR iPhones who want a shareable 3D room without learning 3D tools.

痛点

3D Model Anything needs Xcode to install the capture app, Node and Python locally, and an AI coding agent driving an iterative skill; non-engineers cannot use it.

为何是这个分数

分数较低的原因： Crowded room-scanning market with polished free/cheap apps, and the agent workflow is slow and iterative, not one-click.

护城河假设

Weak: The code is MIT and public. Only the hosted, agent-run UX differs. Status: Hypothesis.

需求信号

X 上的帖子（7 天）

信号趋势

Up from zero 0% Sun→Fri (full UTC days: Sun 18 0 → Fri 23 19); peak Fri 23 19; Fri 23 is 0% below peak

网络需求已验证

GitHub: 1,338 stars in 3 days, rising (above). X: 35 posts in 7 days, rising from 1 to 19/day (Verified X API) — small. A YouTube video by Wes Bos launched it (YouTube is HARK's domain; views not checked).

市场与竞争对手

TAM

Not sized this run — Scope: Consumer/prosumer 3D scanning. Method: No report fetched. (Basis: n/a)

SAM假设

Not sized — Scope: LiDAR iPhone owners with rooms to share. Method: HBV assumption. (Basis: Hypothesis)

SOM假设

~$6K–$36K ARR (hypothesis) — Scope: Month-6 run-rate. Method: From revenue hypothesis ×12. (Basis: Hypothesis)

竞争对手

Polycam; Matterport; magicplan; Luma AI; Apple RoomPlan-based scanner apps (3D Scanner App, Canvas); Zillow 3D Home.

商业模式

变现方式假设

Pay per room (hypothesis ~US$9–19) or Pro around US$15/month for unlimited rooms and branded share links; MIT allows commercial hosting with attribution.

收入假设（每月）假设

~$500–$3K/month at month 6 (HYPOTHESIS: 100–300 rooms/month × ~US$12).

待完成的任务

When I want people to see my space before they visit, I want to turn a quick phone scan into a 3D room they can walk through, so I can share one link and get more bookings.

HXC

Marina, 36, Airbnb host in Florianópolis (illustrative) · Has an iPhone Pro; wants guests to walk through the flat before booking. Buying signal: Would pay per room if it beats Polycam on looks and ease.

构建与 GTM

技术栈

Next.js on Vercel; sandboxed Claude Agent SDK runner; Three.js viewer from the repo (MIT); Vercel Blob/R2 for scans; Stripe

GTM 渠道

Shareable 3D room links (every tour is an ad); real-estate and Airbnb host communities; SEO for "scan room to 3D model free"

风险

主要风险

Crowded room-scanning market with polished free/cheap apps, and the agent workflow is slow and iterative, not one-click.

全部风险已验证

Value: Supported — 1,338 stars in 3 days, rising (Verified) | Usability: High — Crowded scanning apps | Feasibility: Attention — LiDAR-only capture | Business Viability: Attention — Agent cost and time per room

下一步验证

Not recommended now. If tested: run the skill on 3 HBV team rooms to measure minutes and LLM cost per room; kill if over 30 minutes or US$5 per room.

来源

ideanewgithub-wrapper3dlidarhomereal-estatethree-jsglobal

想在评分当天收到下一个想法吗？

分数、来源、风险和下一步，通过邮件发送。

ProductBots.ai

喜欢这个想法吗？用 ProductBots.ai 把它变成计划。

ProductBots.ai 是一个 Agentic Product OS：由智能体与你一起完成产品工作，从发现到计划。下载这个想法的简报，作为起点。

打开 ProductBots.ai

简报文件为英文。

预览简报

# Briefing: 3D-Model-Anything-powered room-to-3D tour web app (working name) — upload an iPhone room scan, get an editable, shareable 3D room you can walk through in the browser

Source: Startup Ideas HQ (HB Ventures), found 2026-10-24. ProductBots Opportunity Score: 51/100, evidence confidence 48%.
Evidence labels (Verified, Self-reported, Hypothesis) are kept as written by the researcher.

## Goal for ProductBots.ai
Turn this idea into a validated MVP plan and a first build. Start from the next validation step; do not assume unlabelled numbers are verified.

## What it is
A hosted, no-terminal version of Wes Bos's 3D Model Anything: scan a room on a LiDAR iPhone, upload the export, and an AI agent rebuilds it as an editable Three.js scene with dollhouse and walk-through views and a share link.

Benchmark: wesbos/3d-model-anything — total 1,338 stars (gh api / OSS Insight, fetched 2026-10-24 07:02 BRT; browser-confirmed (abbreviated 1.3k, repo page 07:19–07:30 BRT), exact from API/OSS Insight); created 2026-10-21 (age 3 days); 24h (23 Oct UTC) +578; 7d (17–23 Oct) +1,046 vs prior 7d (10–16 Oct) +0 = n/a (n/a (no prior)); 3d (21–23 Oct) +1,046 vs prior 3d (18–20 Oct) +0 = n/a (n/a (no prior)); since creation 446.0/day; peak day 2026-10-23 (+578), 0 days since peak; license NOASSERTION; forks 220. Daily stars (OSS Insight): 21 Oct +1, 22 Oct +467, 23 Oct +578, 24 Oct partial +267. Description: "Capture iPhone LiDAR and photos, reconstruct editable Three.js rooms, and explore them in a browser." License: code and skill MIT; office demo images/textures under separate ASSETS-LICENSE (GitHub shows NOASSERTION because of that split) — commercial hosting of the code is allowed with the MIT notice. Method: daily new stars from OSS Insight analyze-recent-stars (last 28 days + prior 28, UTC days) and totals from the GitHub API (gh api repos/OWNER/REPO, cross-check), fetched 24 Oct 2026 ~06:59–07:40 BRT; totals appended to /workspace/hbv_star_snapshots.csv with source. 24h = 23 Oct UTC; 7d = 17–23 Oct vs 10–16 Oct; 3d = 21–23 Oct vs 18–20 Oct; label: >+10% accelerating, <−10% decelerating, otherwise steady; peak over the ~56 days available. Browser check (logged-out github.com, 24 Oct 2026): trending pages read 06:58 BRT (weekly all languages, daily, Python weekly, TypeScript weekly) list none of these repos; repo pages read 07:19–07:30 BRT confirm the totals (browser-confirmed (abbreviated), exact from API/OSS Insight); star-history charts were read for shape only, not values. Star-history shape (browser, approximate axis read, shape only): near-linear steady rise from ~zero on Thu 22 Oct to the current total on Sat 24 Oct; oldest commit 21 Oct 2026; sidebar shows no named license (README: source and skill MIT, separate ASSETS-LICENSE.md). None are estimated. Existence check: Checked 24 Oct 2026: the maintainer (Wes Bos) ships the iPhone app as source only ("an App Store release is a separate distribution step") and does not sell a hosted version; no hosted wrapper of this repo was found in this run's GitHub/web checks (wrapper count: none found; forks 220 are not products). Hosted room scanning already exists from Polycam, Matterport, magicplan, Luma and RoomPlan-based apps. X: 35 posts/7d (Up from zero 0% Sun→Fri (full UTC days: Sun 18 0 → Fri 23 19); peak Fri 23 19; Fri 23 is 0% below peak; window 2026-10-17 10:02–10:04 UTC (start varies by query) → 2026-10-24 ~10:03 UTC (X API 7-day limit; Sat 17 and Sat 24 are partial UTC days; full days Sun 18 → Fri 23)). Revenue hypothesis: ~$500–$3K/month at month 6 (HYPOTHESIS: 100–300 rooms/month × ~US$12).

## Customer and pain
Target customer: Home renovators, Airbnb hosts, small real-estate agents and interior-design hobbyists with LiDAR iPhones who want a shareable 3D room without learning 3D tools.
Pain: 3D Model Anything needs Xcode to install the capture app, Node and Python locally, and an AI coding agent driving an iterative skill; non-engineers cannot use it.

## Market and competitors
TAM: Not sized this run — Scope: Consumer/prosumer 3D scanning. Method: No report fetched. (Basis: n/a)
SAM: Not sized — Scope: LiDAR iPhone owners with rooms to share. Method: HBV assumption. (Basis: Hypothesis)
SOM: ~$6K–$36K ARR (hypothesis) — Scope: Month-6 run-rate. Method: From revenue hypothesis ×12. (Basis: Hypothesis)
Competitors: Polycam; Matterport; magicplan; Luma AI; Apple RoomPlan-based scanner apps (3D Scanner App, Canvas); Zillow 3D Home.
Geography: Global (English first); Portuguese and Spanish later

## Business model
Monetization: Pay per room (hypothesis ~US$9–19) or Pro around US$15/month for unlimited rooms and branded share links; MIT allows commercial hosting with attribution.
Revenue hypothesis (monthly): ~$500–$3K/month at month 6 (HYPOTHESIS: 100–300 rooms/month × ~US$12).

## Build and go-to-market
Build effort (agent-weeks): 3–4 agent-weeks (upload intake for house-capture exports, sandboxed agent runs of the scan-to-threejs skill, viewer hosting, share links, Stripe credits)
Time to first revenue: ~1–2 months
Tech stack: Next.js on Vercel; sandboxed Claude Agent SDK runner; Three.js viewer from the repo (MIT); Vercel Blob/R2 for scans; Stripe
GTM channels: Shareable 3D room links (every tour is an ad); real-estate and Airbnb host communities; SEO for "scan room to 3D model free"

## Risks and next step
Main risk: Crowded room-scanning market with polished free/cheap apps, and the agent workflow is slow and iterative, not one-click.
Next validation action: Not recommended now. If tested: run the skill on 3 HBV team rooms to measure minutes and LLM cost per room; kill if over 30 minutes or US$5 per room.

## Sources
https://github.com/wesbos/3d-model-anything; https://wesbos.github.io/3d-model-anything/; https://ossinsight.io; https://poly.cam; https://matterport.com; https://www.magicplan.app; X API counts (window above)

这是研究，不是投资建议。数字保留其发布时的标签。

摘要

简报○假设≈估算

描述○假设