New ideaOpen source

Carousel Skill

2 weeks ago · Scored 24 Oct 2026 · ID HBV-261024-10

Working title: Carousel Skill — open-source agent skill that turns any app or product into a week of faceless TikTok and Instagram photo-carousel posts (working name)

Summary

BriefingHypothesisEstimate

New idea. Benchmark: Comparable repos (real daily stars): lemomo-ai/lemo-opuscar — total 6,303 stars (gh api / OSS Insight, fetched 2026-10-24 07:04 BRT; browser-confirmed (abbreviated 6.3k, repo page 07:19–07:30 BRT), exact from API/OSS Insight); created 2026-09-26 (age 28 days); 24h (23 Oct UTC) +568; 7d (17–23 Oct) …

Read the rest

+3,379 vs prior 7d (10–16 Oct) +1,160 = +191% (accelerating); 3d (21–23 Oct) +1,775 vs prior 3d (18–20 Oct) +1,407 = +26% (accelerating); since creation 225.1/day; peak day 2026-10-19 (+665), 4 days since peak; license MIT; forks 707. | feitangyuan/onetake — total 3,132 stars (gh api / OSS Insight, fetched 2026-10-24 06:59 BRT; browser-confirmed (abbreviated 3.1k, repo page 07:19–07:30 BRT), exact from API/OSS Insight); created 2026-09-26 (age 28 days); 24h (23 Oct UTC) +107; 7d (17–23 Oct) +560 vs prior 7d (10–16 Oct) +511 = +10% (steady); 3d (21–23 Oct) +328 vs prior 3d (18–20 Oct) +173 = +90% (accelerating); since creation 111.9/day; peak day 2026-09-27 (+365), 26 days since peak; license NOASSERTION; forks 210. | rossark0/rossark-motion-skill — total 409 stars (gh api / OSS Insight, fetched 2026-10-24 07:04 BRT; browser repo page 07:19–07:30 BRT showed 411 exact, live growth since the API read); created 2026-10-22 (age 2 days); 24h (23 Oct UTC) +109; 7d (17–23 Oct) +382 vs prior 7d (10–16 Oct) +0 = n/a (n/a (no prior)); 3d (21–23 Oct) +382 vs prior 3d (18–20 Oct) +0 = n/a (n/a (no prior)); since creation 204.5/day; peak day 2026-10-22 (+273), 1 days since peak; license MIT; forks 31. | athcagithub/SlideSmith — total 338 stars (gh api / OSS Insight, fetched 2026-10-24 07:04 BRT; browser repo page 07:19–07:30 BRT showed 338 exact); created 2026-06-11 (age 135 days); 24h (23 Oct UTC) +1; 7d (17–23 Oct) +3 vs prior 7d (10–16 Oct) +1 = +200% (accelerating); 3d (21–23 Oct) +1 vs prior 3d (18–20 Oct) +2 = -50% (decelerating); since creation 2.5/day; peak day 2026-09-13 (+2), 40 days since peak; license NOASSERTION; forks 77. 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 shapes (browser, approximate axis reads, shape only): lemo-opuscar steady from late Sep, then clearly steeper over the last week; onetake steady, slightly flattening; rossark-motion-skill fast rise Thu 22–Fri 23 Oct, then flattening; SlideSmith nearly flat. GitHub sidebar shows no named license for lemo-opuscar (README: CC BY 4.0 for demo media), onetake and SlideSmith; rossark-motion-skill shows MIT. None are estimated. Existence check: Checked 24 Oct 2026 via GitHub search: athcagithub/SlideSmith (TikTok Slideshow Generator, 338 stars, created 11 Jun 2026, +3 stars in 7 days) and Upload-Post/upload-post-larry-marketing-skill (41 stars) exist; paid SaaS tools for slideshows exist (e.g. ReelFarm-type products, not re-checked). No fast-growing open-source carousel skill was found among trending/new repos. X: 417 posts/7d (Up 98% Sun→Fri (full UTC days: Sun 18 40 → Fri 23 79); peak Thu 22 157; Fri 23 is 50% below peak — ≥50% below peak: POP 5.3 window-closed signal; 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 from the hosted tier (HYPOTHESIS); repo itself earns nothing.

DescriptionHypothesis

A Claude Code / Codex skill: point it at an app URL or repo and it researches hooks, writes 7 days of slideshow scripts, renders 1080×1920 carousel images in code and exports ready-to-post folders — no face, no video editing. Who: Indie app makers and solo founders who market with faceless TikTok/Instagram slideshows (the "slideshow marketing" playbook) and the developers who follow agent-skill repos.…

Read the rest

Pain: Slideshow marketing works for apps but is manual: hooks, image sourcing, text overlay and consistency take hours a day; existing slideshow repos are small and inactive. Monetization: Repo is MIT and free; revenue from a hosted "run it for me" tier with scheduling (~US$19/month, hypothesis) and labeled sponsor slots in the README/site (TrustMRR mechanic). Build: 1–2 agent-weeks (SKILL.md, hook research step, HTML-to-PNG renderer, 10 carousel styles, examples); stack: Claude Code skill (Markdown + scripts); Playwright/Satori PNG rendering; optional Vercel-hosted tier with Stripe. GTM: Launch the repo with faceless example carousels; X dev community; awesome-skills lists; hosted tier linked from README.

Target customer

Indie app makers and solo founders who market with faceless TikTok/Instagram slideshows (the "slideshow marketing" playbook) and the developers who follow agent-skill repos.

Pain

Slideshow marketing works for apps but is manual: hooks, image sourcing, text overlay and consistency take hours a day; existing slideshow repos are small and inactive.

Why this score

Why it scored lower: Repo stars may not convert to paid usage, and TikTok AI-content labelling rules can change.

MoatHypothesis

Moderate early: First good open-source carousel skill can own the stars. Easy to fork later. Status: Hypothesis.

Demand signals

X posts (7 days)

417

Signal trend

Up 98% Sun→Fri (full UTC days: Sun 18 40 → Fri 23 79); peak Thu 22 157; Fri 23 is 50% below peak — ≥50% below peak: POP 5.3 window-closed signal

Web demandVerified

X: slideshow-marketing chatter 417 posts in 7 days, Sun 18 (40) → Fri 23 (79), peak Thu 22 (157) (Verified X API); a TikTok slideshow skill post reached ~24K impressions this week (Verified X API, earlier read). GitHub: agent-made social video repos accelerating (lemo-opuscar +191% WoW).

Market and competitors

TAM

Not sized this run — Scope: Social content automation. Method: No report fetched. (Basis: n/a)

SAMHypothesis

Not sized — Scope: Indie app marketers. Method: HBV assumption. (Basis: Hypothesis)

SOMHypothesis

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

Competitors

SlideSmith (338 stars, dormant); Upload-Post larry skill (41 stars); lemo-opuscar, onetake, video-talkcraft (video, not carousels); paid slideshow SaaS tools; Canva templates.

Business model

MonetizationHypothesis

Repo is MIT and free; revenue from a hosted "run it for me" tier with scheduling (~US$19/month, hypothesis) and labeled sponsor slots in the README/site (TrustMRR mechanic).

Revenue hypothesis (monthly)Hypothesis

~$500–$3K/month at month 6 from the hosted tier (HYPOTHESIS); repo itself earns nothing.

Jobs to be done

When I need daily TikTok posts for my app but hate making them, I want to get a week of on-brand carousels from my app link, so I can grow installs without showing my face.

HXC

Leo, 26, indie app maker in Medellín (illustrative) · Ships small iOS apps; heard slideshows drive installs. Buying signal: Would star the repo and try it; might pay for scheduling.

Build and go-to-market

Tech stack

Claude Code skill (Markdown + scripts); Playwright/Satori PNG rendering; optional Vercel-hosted tier with Stripe

GTM channels

Launch the repo with faceless example carousels; X dev community; awesome-skills lists; hosted tier linked from README

Risks

Main risk

Repo stars may not convert to paid usage, and TikTok AI-content labelling rules can change.

All risksVerified

Value: Supported — Dormant competitors (Verified stars) | Usability: Attention — Spiky X demand | Feasibility: High — No direct revenue | Business Viability: Attention — Platform AI labels

Next validation step

If Cesar wants a GitHub launch: build the skill in 1 week and measure stars in the first 72 hours against rossark-motion-skill (409 in 2 days).

Sources

ideanewgithub-newagent-skilltiktokslideshowsfaceless-marketingopen-sourceglobal

Want the next one the day it is scored?

The score, the sources, the risks and the next step, by email.

ProductBots.ai

Like this idea? Turn it into a plan with ProductBots.ai.

ProductBots.ai is an Agentic Product OS: agents that do the product work with you, from discovery to plan. Download the briefing for this idea and use it as the starting point.

Open ProductBots.ai

Preview the briefing

# Briefing: Carousel Skill — open-source agent skill that turns any app or product into a week of faceless TikTok and Instagram photo-carousel posts (working name)

Source: Startup Ideas HQ (HB Ventures), found 2026-10-24. ProductBots Opportunity Score: 62/100, evidence confidence 51%.
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 Claude Code / Codex skill: point it at an app URL or repo and it researches hooks, writes 7 days of slideshow scripts, renders 1080×1920 carousel images in code and exports ready-to-post folders — no face, no video editing.

New idea. Benchmark: Comparable repos (real daily stars): lemomo-ai/lemo-opuscar — total 6,303 stars (gh api / OSS Insight, fetched 2026-10-24 07:04 BRT; browser-confirmed (abbreviated 6.3k, repo page 07:19–07:30 BRT), exact from API/OSS Insight); created 2026-09-26 (age 28 days); 24h (23 Oct UTC) +568; 7d (17–23 Oct) +3,379 vs prior 7d (10–16 Oct) +1,160 = +191% (accelerating); 3d (21–23 Oct) +1,775 vs prior 3d (18–20 Oct) +1,407 = +26% (accelerating); since creation 225.1/day; peak day 2026-10-19 (+665), 4 days since peak; license MIT; forks 707. | feitangyuan/onetake — total 3,132 stars (gh api / OSS Insight, fetched 2026-10-24 06:59 BRT; browser-confirmed (abbreviated 3.1k, repo page 07:19–07:30 BRT), exact from API/OSS Insight); created 2026-09-26 (age 28 days); 24h (23 Oct UTC) +107; 7d (17–23 Oct) +560 vs prior 7d (10–16 Oct) +511 = +10% (steady); 3d (21–23 Oct) +328 vs prior 3d (18–20 Oct) +173 = +90% (accelerating); since creation 111.9/day; peak day 2026-09-27 (+365), 26 days since peak; license NOASSERTION; forks 210. | rossark0/rossark-motion-skill — total 409 stars (gh api / OSS Insight, fetched 2026-10-24 07:04 BRT; browser repo page 07:19–07:30 BRT showed 411 exact, live growth since the API read); created 2026-10-22 (age 2 days); 24h (23 Oct UTC) +109; 7d (17–23 Oct) +382 vs prior 7d (10–16 Oct) +0 = n/a (n/a (no prior)); 3d (21–23 Oct) +382 vs prior 3d (18–20 Oct) +0 = n/a (n/a (no prior)); since creation 204.5/day; peak day 2026-10-22 (+273), 1 days since peak; license MIT; forks 31. | athcagithub/SlideSmith — total 338 stars (gh api / OSS Insight, fetched 2026-10-24 07:04 BRT; browser repo page 07:19–07:30 BRT showed 338 exact); created 2026-06-11 (age 135 days); 24h (23 Oct UTC) +1; 7d (17–23 Oct) +3 vs prior 7d (10–16 Oct) +1 = +200% (accelerating); 3d (21–23 Oct) +1 vs prior 3d (18–20 Oct) +2 = -50% (decelerating); since creation 2.5/day; peak day 2026-09-13 (+2), 40 days since peak; license NOASSERTION; forks 77. 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 shapes (browser, approximate axis reads, shape only): lemo-opuscar steady from late Sep, then clearly steeper over the last week; onetake steady, slightly flattening; rossark-motion-skill fast rise Thu 22–Fri 23 Oct, then flattening; SlideSmith nearly flat. GitHub sidebar shows no named license for lemo-opuscar (README: CC BY 4.0 for demo media), onetake and SlideSmith; rossark-motion-skill shows MIT. None are estimated. Existence check: Checked 24 Oct 2026 via GitHub search: athcagithub/SlideSmith (TikTok Slideshow Generator, 338 stars, created 11 Jun 2026, +3 stars in 7 days) and Upload-Post/upload-post-larry-marketing-skill (41 stars) exist; paid SaaS tools for slideshows exist (e.g. ReelFarm-type products, not re-checked). No fast-growing open-source carousel skill was found among trending/new repos. X: 417 posts/7d (Up 98% Sun→Fri (full UTC days: Sun 18 40 → Fri 23 79); peak Thu 22 157; Fri 23 is 50% below peak — ≥50% below peak: POP 5.3 window-closed signal; 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 from the hosted tier (HYPOTHESIS); repo itself earns nothing.

## Customer and pain
Target customer: Indie app makers and solo founders who market with faceless TikTok/Instagram slideshows (the "slideshow marketing" playbook) and the developers who follow agent-skill repos.
Pain: Slideshow marketing works for apps but is manual: hooks, image sourcing, text overlay and consistency take hours a day; existing slideshow repos are small and inactive.

## Market and competitors
TAM: Not sized this run — Scope: Social content automation. Method: No report fetched. (Basis: n/a)
SAM: Not sized — Scope: Indie app marketers. Method: HBV assumption. (Basis: Hypothesis)
SOM: ~$6K–$36K ARR (hypothesis) — Scope: Month-6 run-rate of hosted tier. Method: From revenue hypothesis ×12. (Basis: Hypothesis)
Competitors: SlideSmith (338 stars, dormant); Upload-Post larry skill (41 stars); lemo-opuscar, onetake, video-talkcraft (video, not carousels); paid slideshow SaaS tools; Canva templates.
Geography: Global (English); Portuguese examples

## Business model
Monetization: Repo is MIT and free; revenue from a hosted "run it for me" tier with scheduling (~US$19/month, hypothesis) and labeled sponsor slots in the README/site (TrustMRR mechanic).
Revenue hypothesis (monthly): ~$500–$3K/month at month 6 from the hosted tier (HYPOTHESIS); repo itself earns nothing.

## Build and go-to-market
Build effort (agent-weeks): 1–2 agent-weeks (SKILL.md, hook research step, HTML-to-PNG renderer, 10 carousel styles, examples)
Time to first revenue: ~1–2 months (after stars, via hosted tier)
Tech stack: Claude Code skill (Markdown + scripts); Playwright/Satori PNG rendering; optional Vercel-hosted tier with Stripe
GTM channels: Launch the repo with faceless example carousels; X dev community; awesome-skills lists; hosted tier linked from README

## Risks and next step
Main risk: Repo stars may not convert to paid usage, and TikTok AI-content labelling rules can change.
Next validation action: If Cesar wants a GitHub launch: build the skill in 1 week and measure stars in the first 72 hours against rossark-motion-skill (409 in 2 days).

## Sources
https://github.com/athcagithub/SlideSmith; https://github.com/Upload-Post/upload-post-larry-marketing-skill; https://github.com/lemomo-ai/lemo-opuscar; https://github.com/feitangyuan/onetake; https://github.com/rossark0/rossark-motion-skill; https://ossinsight.io; X API counts (window above)

New ideaOpen source· 3 weeks ago
Shortcraft — an open-source engine that turns…
A new HBV GitHub repo: give it a script or topic and it renders a 9:16 faceless short in one of dozens of code-made styles (kinetic type, diagrams, retro film), with captions and voice, ready for TikTok, Reels and Shorts — no camera, no stock footage.
55/100
confidence 50%
Why it scored lower: HeyGen or lemo-opuscar ship a vertical-shorts preset and the repo has no reason to exist.
Open ProductBots.ai →

Research, not investment advice. Numbers keep the label they were published with.