Przejdź do treści
Biblioteka skilli
Nauka i badania · Research

youtube-transcript-research

/youtube-transcript-research

Bada temat na podstawie transkrypcji filmów YouTube i pisze cytowane podsumowanie z odnośnikami do konkretnych sekund.

Repozytorium
skills ↗
Licencja
MIT
Ostatnia zmiana repozytorium
03.11.2026

Ocena i bezpieczeństwo

100/ 100Klasa: dobry
Format25 / 25
Frontmatter z nazwą i opisem, poprawna nazwa, opis do 1024 znaków, nazwa zgodna z folderem.
Odkrywalność25 / 25
Opis mówi, kiedy użyć skilla, ma sensowną długość i jest w 3. osobie.
Treść35 / 35
Co najmniej 150 słów i do 500 linii, nagłówki, listy kroków, przykłady albo kod.
Higiena15 / 15
Bez zaślepek typu TODO, bez ścieżek Windows, z licencją.
Ryzykobrak sygnałów w skanie
Pliki w paczce
1
Skrypty
0
Komentarze HTML
2
Adresy URL w SKILL.md
5
Przestarzałość

Brak sygnałów. Repozytorium zmienione 03.11.2026, bez odwołań do wycofanych modeli i API.

Licencja

MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.

Co sprawdzić przed instalacją
  • SKILL.md zawiera komentarze HTML (2). Nie widać ich w podglądzie Markdown, a model je czyta.

Co robi i kiedy

Bada temat na podstawie transkrypcji filmów YouTube i pisze cytowane podsumowanie z odnośnikami do konkretnych sekund.

Oryginalny opis (EN)

Research a topic from YouTube videos - search YouTube or list a channel, pull full transcripts with timestamps, and write a cited summary of what the videos claim, where they agree and disagree, with links that jump to the exact second. Use when the user asks what YouTubers say about a topic, wants a YouTube transcript, a video summary, competitor or creator content research, product review roundups from video, or quotes with timestamps. Works when yt-dlp or a transcript API is blocked. Runs Arcadium Data actors on Apify with the user's own APIFY_TOKEN (pay per video).

Jak działa

  1. Example prompts
  2. Prerequisites
  3. Cost
  4. Workflow
  5. MCP instead of the CLI
  6. Troubleshooting
Sekcje z nagłówków SKILL.md, w oryginale · 87 linii, 642 słów
Podgląd SKILL.md
---
name: youtube-transcript-research
description: Research a topic from YouTube videos - search YouTube or list a channel, pull full transcripts with timestamps, and write a cited summary of what the videos claim, where they agree and disagree, with links that jump to the exact second. Use when the user asks what YouTubers say about a topic, wants a YouTube transcript, a video summary, competitor or creator content research, product review roundups from video, or quotes with timestamps. Works when yt-dlp or a transcript API is blocked. Runs Arcadium Data actors on Apify with the user's own APIFY_TOKEN (pay per video).
license: MIT
metadata:
  author: Arcadium Data
  author_url: https://arcadiumdata.io
  category: content
  keywords: "youtube transcript, youtube captions, subtitles, video research, youtube search, channel videos, content research, video summary, timestamps"
  apify-actors: "arcadium/youtube-videos, arcadium/youtube-transcripts"
---

# YouTube transcript research

Find the videos that matter on a topic, read their transcripts, and answer with claims tied to timestamped links.

**Disclosure:** the actors this skill runs are built and sold by Arcadium Data, the author of this skill. You pay Apify per video delivered (see Cost). No affiliate codes are used.

## Example prompts

- "What do the top YouTube reviews of the Framework Laptop 13 complain about? Cite timestamps."
- "Summarise the last 20 videos of @kurzgesagt and list the topics."
- "Get the transcript of https://www.youtube.com/watch?v=dQw4w9WgXcQ."

Out of scope:

- YouTube comments (no public Arcadium actor for them yet).
- Downloading video or audio files.

## Prerequisites

- Apify account and `APIFY_TOKEN`, or `apify login`. Apify CLI (`npm i -g apify-cli`) or the Apify MCP server (`https://mcp.apify.com`).

## Cost

Free Apify plan prices; paid plans pay less. Videos without captions, private, age-restricted or missing videos, repeats and error rows are free.

| Actor | Charged event | Free plan | Start fee per run |
| --- | --- | --- | --- |
| `arcadium/youtube-videos` | `video-listing` | $1.17 per 1,000 | $0.001 |
| `arcadium/youtube-videos` | `video` | $3.60 per 1,000 | $0.001 |
| `arcadium/youtube-transcripts` | `transcript` | $2.48 per 1,000 | $0.001 |

`video-listing` is the fast mode (`listingOnly: true`): what the search or channel listing shows. `video` adds exact date, likes, category, description and caption languages.

Worked example: 30 search results in fast mode, then 15 transcripts = 30 x $0.00117 + $0.001 + 15 x $0.00248 + $0.001 = **about $0.07**.

## Workflow

### 1. Find videos (skip if the user gave links)

<!-- input: youtube-videos -->
```json
{
  "searchQueries": ["framework laptop 13 review"],
  "maxResults": 30,
  "listingOnly": true,
  "dateFilter": "year"
}
```

    apify actors call arcadium/youtube-videos --input-file search.json --user-agent arcadium-skills/youtube-transcript-research --json 2>/dev/null
    apify datasets get-items DATASET_ID --format json --user-agent arcadium-skills/youtube-transcript-research 2>/dev/null > videos.json

(`DATASET_ID` is `.defaultDatasetId` from the call output.) Rows carry `videoId`, `url`, `title`, `channelName`, `viewCount` (rounded in fast mode), `durationSeconds`, `publishDate` (approximate when `publishDateIsApprox` is true) and `position` (search rank). For a channel, put its URL (`https://www.youtube.com/@handle`) in `urls` instead of `searchQueries`.

Pick 10-20 videos: drop Shorts and videos under 3 minutes for in-depth topics, prefer a spread of channels, and tell the user which ones you chose and why.

### 2. Pull the transcripts

<!-- input: youtube-transcripts -->
```json
{
  "urls": ["https://www.youtube.com/watch?v=dQw4w9WgXcQ"],
  "languages": ["en"],
  "includeSegments": true
}
```

    apify actors call arcadium/youtube-transcripts --input-file transcripts.json --user-agent arcadium-skills/youtube-transcript-research --json 2>/dev/null

Put all chosen video URLs in `urls` (one run, one start fee). A channel or playlist URL also works, with `maxVideosPerSource` and `publishedAfter` (`YYYY-MM-DD`) to limit it.

Read from `rowType: "transcript"` rows: `transcript` (full text), `segments` (`start`, `duration` in seconds, `text`), `language`, `isAutoGenerated`, `languageMatched`, `availableLanguages`, plus `title`, `channelName`, `publishDate`, `viewCount`, `wordCount`. For a non-English video, use the native track the video has (`languages`) and translate the text yourself: `translateTo` returns a native track in that language when one exists, and YouTube currently refuses its machine translation to anonymous clients (`translationStatus: "unavailable"`).

### 3. Write the research note

For each video: title, channel, date, views, and 3-6 key claims. For every claim give a timestamp link `<url>&t=<start>s`, using the `start` of the segment where the claim is made (rounded down). Then a cross-video section: points most videos agree on, points of disagreement, and claims only one creator makes. Quote transcript text exactly; mark auto-generated captions (`isAutoGenerated: true`) because names and numbers in them can be misheard.

## MCP instead of the CLI

With the Apify MCP server connected, call `arcadium/youtube-videos` and `arcadium/youtube-transcripts` with the same inputs.

## Troubleshooting

- A free row with `rowType: "error"` and a `code` means no transcript (no captions, private, age-restricted or missing). Check `availableLanguages` on delivered rows to request another language.
- `languageMatched: false`: none of the requested languages exists; the original spoken language was returned.

Wersje

W bazie nie ma innych skilli o nazwie youtube-transcript-research.

Migawka bazy z 13.11.2026 · wersja paczki 551824104729 (skrót SHA-256 całej paczki)

Pobrania w ostatnich 30 dniach: 0 · ranking

Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.

Podobne skille

  • 100youtube-transcript
    Pobiera transkrypcje filmów, Shorts, kanałów i playlist YouTube jako tekst, segmenty, SRT lub VTT, z metadanymi.
  • 100youtube-transcript
    Pobiera transkrypcję filmu YouTube, gdy użytkownik prosi o tekst, napisy lub streszczenie treści z linku.
  • 100video-transcript-research
    Pobiera i analizuje transkrypcje wideo z YouTube, TikToka, Instagrama, X i LinkedIn, porównując wypowiedzi wielu twórców.
  • 100okdata-transcripts
    Pobiera transkrypcję filmu z YouTube z sygnaturami czasowymi i przygotowuje streszczenie, cytaty lub porównanie kilku filmów.
  • 100yt-transcript
    Pobiera tekst wypowiedzi z filmów YouTube w postaci napisów lub transkrypcji, także do dalszego przetwarzania treści wideo.
  • 90youtube-transcript
    Pobiera transkrypcję z linku lub ID filmu YouTube i przygotowuje podsumowanie, notatki, cytaty, znaczniki czasu lub tłumaczenie.

Uwagi i zgłoszenia

Wiadomość trafi do Bartosza Sokolińskiego. Adres e-mail jest potrzebny tylko do odpowiedzi. Szczegóły w polityce prywatności.