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Creator tools

rea-video-powered ad teardown

A hosted web app on top of the open-source rea-video engine: paste a winning TikTok/Reels ad (or upload the file) and get a shot-by-shot report of what was filmed, AI-generated or coded, how each shot was lit and edited, a reverse-engineered prompt per AI shot, and a remake plan with your own product.

1周前 · 评分于 2026年11月5日 · ID HBV-261105-01

暂定标题： rea-video-powered ad teardown for TikTok Shop and Reels sellers — paste an ad link, get shot-by-shot how it was made and a remake plan (Portuguese/Spanish/English) — working name "Desmonta"

摘要

简报已验证自行申报假设估算

Benchmark: JuneYaooo/rea-video: 1,603 stars in 7 days (+806 on 4 Nov, peak), Apache-2.0, 232 forks, runs in Claude Code/Codex, optional Gemini/Doubao video model; README accuracy table 99.6% with a vision model (Self-reported) (stars Verified, github.com and OSS Insight). Existence check: Paid ad-creative tools exist (…

阅读其余部分

Foreplay, Motion, Atria, Kalodata/FastMoss ad libraries) but none found that tells shot by shot whether footage was filmed or AI-generated and writes a remake plan; open-source Hypit (hypit-ai/hypit) remakes videos but does not explain how the original was made (README; quick search, Hypothesis). POS 63 is on the line but not strictly above 70: crowded ad-creative-analysis category (Foreplay, Motion, Atria, Kalodata/FastMoss) and the X wave is only 2 days old; the need for the user's own coding agent is only a risk, not a cut. Stars: Total 1,603 (github.com HTML 5 Nov ~07:06 BRT; API 1,603). Created 29 Oct 2026 (7 days), Apache-2.0, 232 forks. 24h (4 Nov) +806; 7d (29 Oct–4 Nov) +941 ≈134/day; prior 7d 0 (repo did not exist) so 7v7 n/a; 3v3 (2–4 Nov 941 vs 30 Oct–1 Nov 0) n/a, day over day +24 → +111 → +806 (accelerating); peak 4 Nov (+806), 0 days since peak; since creation ≈229/day; today partial +650 (to ~07:04 BRT). Nothing estimated. X: 453 posts/7d (Rising +3712% 3v3 (2–4 Nov 305 vs 30 Oct–1 Nov 8; daily full UTC days 5, 2, 1, 2, 25, 278); peak 4 Nov 278; 4 Nov is 0% below peak). Revenue hypothesis: ~$1K–$5K/month at month 6 (HYPOTHESIS).

描述假设

阅读其余部分

Who: TikTok Shop, Shopee and Mercado Livre sellers, UGC creators and small performance agencies in LatAm and globally who copy winning ad formats every week. Pain: Copying a winning ad today means watching it frame by frame and guessing whether it was filmed or AI-generated; rea-video does this but needs Python, ffmpeg and a coding agent (Claude Code/Codex) on your own machine. Monetization: Freemium: 3 free teardowns, then monthly plans with a teardown quota; agency plan with team folders (Hypothesis; no approved price). GTM: SEO/GEO pages per viral ad format ('how this TikTok ad was made'), public teardown gallery as a PLG loop (every report has a share link), faceless short clips of teardowns, LatAm seller communities.

目标客户

TikTok Shop, Shopee and Mercado Livre sellers, UGC creators and small performance agencies in LatAm and globally who copy winning ad formats every week.

痛点

Copying a winning ad today means watching it frame by frame and guessing whether it was filmed or AI-generated; rea-video does this but needs Python, ffmpeg and a coding agent (Claude Code/Codex) on your own machine.

为何是这个分数

分数较低的原因： Crowded paid ad-creative category and free general video models (Gemini, ChatGPT) cover part of the job; the X wave is only 2 days old and could fade.

护城河假设

Weak to moderate: Public teardown gallery compounds SEO. Portuguese/Spanish first for LatAm sellers. Status: Hypothesis.

需求信号

X 上的帖子（7 天）

453

信号趋势

Rising +3712% 3v3 (2–4 Nov 305 vs 30 Oct–1 Nov 8; daily full UTC days 5, 2, 1, 2, 25, 278); peak 4 Nov 278; 4 Nov is 0% below peak

网络需求自行申报

README with six English example reports (TikTok ad, AI short drama, Remotion motion graphics); accuracy table 99.6–99.8% with a Gemini vision check (Self-reported by maintainer).

市场与竞争对手

TAM

Not sized this run — Scope: Global. Method: no third-party report fetched this run (token saving). (Basis: n/a)

SAM假设

Not sized — Scope: Target markets below. Method: HBV assumption. (Basis: Hypothesis)

SOM假设

~$12K–$60K ARR (Hypothesis) — Scope: Month-6 run-rate. Method: Revenue hypothesis ×12. (Basis: Hypothesis)

竞争对手

Foreplay, Motion (creative analytics), Atria, Kalodata/FastMoss (TikTok Shop ad libraries), Meta Ad Library (free), Gemini/ChatGPT video upload (free/partial), Hypit (open source remake).

商业模式

变现方式假设

Freemium: 3 free teardowns, then monthly plans with a teardown quota; agency plan with team folders (Hypothesis; no approved price).

收入假设（每月）假设

~$1K–$5K/month at month 6 (HYPOTHESIS).

待完成的任务

When I see a competitor's ad going viral, I want to know shot by shot how it was made and what to film or generate, so I can remake it with my product the same day.

HXC

Camila, 31, TikTok Shop Brazil seller in São Paulo (illustrative) · Sells beauty products, posts 5 ads a week, copies formats that go viral. Buying signal: Pays for an ad library and AI video credits.

构建与 GTM

技术栈

未知

GTM 渠道

SEO/GEO pages per viral ad format ('how this TikTok ad was made'), public teardown gallery as a PLG loop (every report has a share link), faceless short clips of teardowns, LatAm seller communities

风险

主要风险

Crowded paid ad-creative category and free general video models (Gemini, ChatGPT) cover part of the job; the X wave is only 2 days old and could fade.

全部风险

Value: Attention — Willingness to pay vs free ad libraries | Usability: Attention — Platform terms for fetching videos | Feasibility: Supported — Apache CLI already works | Business Viability: High — Wave 2 days old; crowded category

下一步验证

Run 5 real TikTok Shop Brazil ads through rea-video + Gemini and measure cost and time per teardown; ask HBV's TikTok Shopping business whether it would use it; re-check stars and X on 6 Nov.

来源

github.com
github.com
OSS Insight analyze-recent-stars
X API counts (window above)

newgithub-wrapperborderlinegithubcreator-toolsecommercetiktok-shoplatamglobal

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: rea-video-powered ad teardown for TikTok Shop and Reels sellers — paste an ad link, get shot-by-shot how it was made and a remake plan (Portuguese/Spanish/English) — working name "Desmonta"

Source: Startup Ideas HQ (HB Ventures), found 2026-11-05. ProductBots Opportunity Score: 63/100, evidence confidence 53%.
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 web app on top of the open-source rea-video engine: paste a winning TikTok/Reels ad (or upload the file) and get a shot-by-shot report of what was filmed, AI-generated or coded, how each shot was lit and edited, a reverse-engineered prompt per AI shot, and a remake plan with your own product.

Benchmark: JuneYaooo/rea-video: 1,603 stars in 7 days (+806 on 4 Nov, peak), Apache-2.0, 232 forks, runs in Claude Code/Codex, optional Gemini/Doubao video model; README accuracy table 99.6% with a vision model (Self-reported) (stars Verified, github.com and OSS Insight). Existence check: Paid ad-creative tools exist (Foreplay, Motion, Atria, Kalodata/FastMoss ad libraries) but none found that tells shot by shot whether footage was filmed or AI-generated and writes a remake plan; open-source Hypit (hypit-ai/hypit) remakes videos but does not explain how the original was made (README; quick search, Hypothesis). POS 63 is on the line but not strictly above 70: crowded ad-creative-analysis category (Foreplay, Motion, Atria, Kalodata/FastMoss) and the X wave is only 2 days old; the need for the user's own coding agent is only a risk, not a cut. Stars: Total 1,603 (github.com HTML 5 Nov ~07:06 BRT; API 1,603). Created 29 Oct 2026 (7 days), Apache-2.0, 232 forks. 24h (4 Nov) +806; 7d (29 Oct–4 Nov) +941 ≈134/day; prior 7d 0 (repo did not exist) so 7v7 n/a; 3v3 (2–4 Nov 941 vs 30 Oct–1 Nov 0) n/a, day over day +24 → +111 → +806 (accelerating); peak 4 Nov (+806), 0 days since peak; since creation ≈229/day; today partial +650 (to ~07:04 BRT). Nothing estimated. X: 453 posts/7d (Rising +3712% 3v3 (2–4 Nov 305 vs 30 Oct–1 Nov 8; daily full UTC days 5, 2, 1, 2, 25, 278); peak 4 Nov 278; 4 Nov is 0% below peak). Revenue hypothesis: ~$1K–$5K/month at month 6 (HYPOTHESIS).

## Customer and pain
Target customer: TikTok Shop, Shopee and Mercado Livre sellers, UGC creators and small performance agencies in LatAm and globally who copy winning ad formats every week.
Pain: Copying a winning ad today means watching it frame by frame and guessing whether it was filmed or AI-generated; rea-video does this but needs Python, ffmpeg and a coding agent (Claude Code/Codex) on your own machine.

## Market and competitors
TAM: Not sized this run — Scope: Global. Method: no third-party report fetched this run (token saving). (Basis: n/a)
SAM: Not sized — Scope: Target markets below. Method: HBV assumption. (Basis: Hypothesis)
SOM: ~$12K–$60K ARR (Hypothesis) — Scope: Month-6 run-rate. Method: Revenue hypothesis ×12. (Basis: Hypothesis)
Competitors: Foreplay, Motion (creative analytics), Atria, Kalodata/FastMoss (TikTok Shop ad libraries), Meta Ad Library (free), Gemini/ChatGPT video upload (free/partial), Hypit (open source remake).
Geography: Global (English) plus LatAm versions in Portuguese and Spanish (TikTok Shop Brazil/Mexico, Mercado Livre sellers)

## Business model
Monetization: Freemium: 3 free teardowns, then monthly plans with a teardown quota; agency plan with team folders (Hypothesis; no approved price).
Revenue hypothesis (monthly): ~$1K–$5K/month at month 6 (HYPOTHESIS).

## Build and go-to-market
Build effort (agent-weeks): ~2–3 agent-weeks (upload/link intakeworker running rea-video CLI with ffmpeg, Gemini video calls, report page, Stripe)
Time to first revenue: ~3–5 weeks
Tech stack: unknown
GTM channels: SEO/GEO pages per viral ad format ('how this TikTok ad was made'), public teardown gallery as a PLG loop (every report has a share link), faceless short clips of teardowns, LatAm seller communities

## Risks and next step
Main risk: Crowded paid ad-creative category and free general video models (Gemini, ChatGPT) cover part of the job; the X wave is only 2 days old and could fade.
Next validation action: Run 5 real TikTok Shop Brazil ads through rea-video + Gemini and measure cost and time per teardown; ask HBV's TikTok Shopping business whether it would use it; re-check stars and X on 6 Nov.

## Sources
https://github.com/JuneYaooo/rea-video; https://github.com/hypit-ai/hypit; OSS Insight analyze-recent-stars; X API counts (window above)

新想法Creator tools· 1周前
Talking-avatar / lip-sync web studio
SaaS/API talking-avatar studio (upload face/audio → embeddable talking clip or live web avatar).
60/100
可信度 50%
分数较低的原因： Free OSS competition: Apache-2.0 player with 877 stars in 2 days; featured characters non-commercial only.
打开 ProductBots.ai →
新想法Creator tools· 1周前
Hosted AI watermark stripper for creators
Self-serve API and web app: upload AI images or short videos you own, strip or normalize SynthID-style watermarks and C2PA manifests in batch, get webhooks and usage billing.
57/100
可信度 50%
分数较低的原因： Legal/ToS risk and free OSS kill willingness to pay; platforms may punish stripped provenance.
打开 ProductBots.ai →
新想法Creator tools· 1周前
Hosted pixel-art effect studio and API
Web app and API: upload an image or short clip, apply a library of code-defined pixel effects, export PNG/GIF/MP4 or an effect pack for games and sticker packs.
57/100
可信度 52%
分数较低的原因： Free desktop tools and a cooling GitHub wave.
打开 ProductBots.ai →

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✓已验证◐自行申报○假设≈估算

描述○假设

目标客户

痛点

为何是这个分数

护城河○假设

需求信号

X 上的帖子（7 天）

信号趋势

网络需求◐自行申报

市场与竞争对手

TAM

SAM○假设

SOM○假设