Youtube Insight Extractor

What does YouTube Insight Extractor do?

YouTube Insight Extractor reads a YouTube video's transcript and returns a structured briefing for each video: a short summary, key takeaways, action items, notable quotes with timestamps, an ordered chapter breakdown, topic tags, tone, and target audience. It also returns the full transcript, the timestamped transcript segments, and the video's metadata (title, channel, duration, view count, description, keywords, thumbnail). You give it a list of YouTube video URLs or IDs; it gives you one clean row per video that you can read, search, or feed into notes and research tools.

It is built for researchers, content teams, analysts, and anyone who needs the substance of a video without watching all of it.

Why use YouTube Insight Extractor?

• Substance, not a raw dump - you get takeaways and action items, not just an unformatted transcript.
• Timestamped - notable quotes and chapters carry offsets so you can jump straight to the moment.
• Full transcript included - the complete transcript and per-line segments are in every row for your own processing.
• Clean, structured output - one row per video, consistent field names, ready for CSV or API.
• Any available caption language - request a language, or fall back to the video's default track.
• Automatable - schedule runs over a channel's new uploads or a research list.

What makes this different

Most transcript tools return the transcript and stop. This Actor adds the briefing layer: it turns the transcript into a summary, a list of concrete takeaways, action items a viewer could act on, verbatim quotes with timestamps, and a chapter outline - while still returning the raw transcript and segments so nothing is lost. The briefing uses a no-cost default model, so the per-result cost stays low.

How it works

1. Your input - a list of YouTube video URLs or IDs and a preferred transcript language.
2. The Actor collects it - it reads each video's metadata and transcript, then writes one structured briefing per video.
3. Your output - a dataset with one row per video: summary, takeaways, action items, quotes, chapters, topics, and the full transcript. Export to JSON, CSV, or Excel, or read it from the API.

What data can you extract?

Field	Description
videoId	The 11-character YouTube video ID.
url	Canonical watch URL.
title	Video title.
channel	Channel name.
channelUrl	Channel URL.
publishDate	Publish or upload date, when available.
durationSeconds	Video length in seconds.
viewCount	View count at read time.
category	YouTube category, when available.
keywords	Video keyword tags.
description	Video description text.
thumbnailUrl	Highest-resolution thumbnail URL.
transcriptLanguage	Language code of the transcript used.
transcriptStatus	ok, no-captions, or a short reason.
summary	3-6 sentence summary of the video.
keyTakeaways	List of concise takeaway points.
actionItems	Concrete things a viewer could do; empty for non-instructional videos.
topics	Topic tags derived from the transcript.
notableQuotes	List of { quote, approxOffsetSeconds, timestamp }.
chapters	Ordered list of { title, summary, startOffsetSeconds, timestamp }.
tone	The video's tone.
targetAudience	Who the video is aimed at.
transcript	The full transcript as one string.
transcriptSegments	Per-line { text, offsetSeconds, timestamp }.
enrichmentConfidence	Briefing confidence: high, medium, or low.
enrichmentStatus	Whether the briefing was produced, and if not, why.
scrapedAt	ISO timestamp of the run.

How to use YouTube Insight Extractor

1. Open the Actor in the Apify Console.
2. In YouTube videos, paste one video URL or ID per line.
3. Set Max videos and, if needed, a Preferred transcript language.
4. Press Start.
5. When the run finishes, open the Dataset tab and export to CSV, JSON, or Excel, or connect it to your workflow.

Input Parameters

Parameter	Type	Required	Default	Description
videoUrls	array	Yes	one example video	One entry per video: a watch URL, youtu.be link, Shorts link, or bare video ID.
maxItems	integer	No	5	Stop after this many video reports.
language	string	No	en	Two-letter language code for the transcript. Falls back to the default track if unavailable.
openRouterApiKey	string (secret)	No	—	Optional key for the analysis service. Without it, the transcript and metadata are still returned and the briefing fields are left empty.
proxyConfiguration	object	Yes	Apify residential	Proxy settings. A residential proxy is required; the default uses Apify's residential proxy.

Output Data

One dataset item per video. Real sample record (transcript trimmed for length):

{
    "videoId": "UF8uR6Z6KLc",
    "url": "https://www.youtube.com/watch?v=UF8uR6Z6KLc",
    "title": "Steve Jobs' 2005 Stanford Commencement Address",
    "channel": "Stanford",
    "channelUrl": "https://www.youtube.com/channel/UC-EnprmCZ3OXyAoG7vjVNCA",
    "publishDate": "2008-03-07",
    "durationSeconds": 904,
    "viewCount": 48969661,
    "category": "Education",
    "keywords": ["apple", "graduation", "stanford", "commencement", "Steve", "Jobs"],
    "description": "Drawing from some of the most pivotal points in his life, Steve Jobs...",
    "thumbnailUrl": "https://i.ytimg.com/vi/UF8uR6Z6KLc/hqdefault.jpg",
    "transcriptLanguage": "en",
    "transcriptStatus": "ok",
    "summary": "Steve Jobs' 2005 Stanford Commencement Address shares three personal stories illustrating the importance of connecting the dots, following one's heart, and embracing death as life's change agent. He emphasizes the significance of loving one's work and staying hungry and foolish.",
    "keyTakeaways": [
        "Trust that the dots will connect in your future, even when you can't see how.",
        "Love what you do, and don't settle until you find it.",
        "Embrace the fact that death is life's change agent and use it as a motivator to live authentically.",
        "Stay hungry and foolish in your pursuits."
    ],
    "actionItems": [
        "Reflect on your life and choices, trusting that they will connect in the future.",
        "Pursue your passions and love what you do."
    ],
    "topics": ["Connecting the dots", "Love and loss", "Death and life", "Following one's heart"],
    "notableQuotes": [
        {
            "quote": "You can't connect the dots looking forward; you can only connect them looking backwards.",
            "approxOffsetSeconds": 120,
            "timestamp": "2:00"
        },
        { "quote": "Stay Hungry. Stay Foolish.", "approxOffsetSeconds": 1080, "timestamp": "18:00" }
    ],
    "chapters": [
        {
            "title": "Connecting the Dots",
            "summary": "Jobs shares how dropping out of Reed College led him to a calligraphy class that later influenced the Macintosh.",
            "startOffsetSeconds": 0,
            "timestamp": "0:00"
        },
        {
            "title": "Love and Loss",
            "summary": "Jobs discusses being fired from Apple and how it led to NeXT and Pixar.",
            "startOffsetSeconds": 300,
            "timestamp": "5:00"
        }
    ],
    "tone": "Inspirational, reflective, and motivational",
    "targetAudience": "Graduates and those seeking inspiration in life and career",
    "transcript": "This program is brought to you by Stanford University... I am honored to be with you today...",
    "transcriptSegments": [
        { "text": "This program is brought to you by Stanford University.", "offsetSeconds": 7, "timestamp": "0:07" }
    ],
    "enrichmentConfidence": "high",
    "enrichmentStatus": "ok",
    "scrapedAt": "2026-09-04T11:11:22.157Z"
}

Usage Examples

• Research digest - paste 20 talk URLs, run, and read the summary and keyTakeaways columns.
• Course notes - run a tutorial series and keep the chapters and actionItems for each video.
• Quote sourcing - filter notableQuotes for a phrase and jump to the timestamp.
• Content repurposing - use topics and keyTakeaways as the outline for a blog post or newsletter.

Tips for Best Results

• Videos need captions (creator-provided or auto-generated). transcriptStatus tells you when a video has none.
• Set language to match the spoken language for the cleanest transcript.
• Long videos are summarized from the first portion of the transcript when it is very long; the full transcript is always returned in full.
• Provide an analysis service API key for the briefing fields; the transcript and metadata work without one.

Known Limitations

• A residential proxy is required; video data is not served to datacenter traffic.
• Videos with captions disabled return transcriptStatus no-captions and empty briefing fields.
• Private, age-restricted, members-only, and region-blocked videos may return a status explaining why instead of a transcript.
• publishDate and category depend on the video page and can be null.
• Auto-generated captions contain transcription errors that carry into the briefing.
• The briefing quality reflects the default no-cost model and can be terse on very long or highly technical videos.

Integrations

Connect the dataset to Google Sheets, Airtable, Make, Zapier, or n8n; call the Actor from the Apify API; trigger it with webhooks; or schedule recurring runs from the Apify Console.

Export Formats

JSON, CSV, Excel, and XML, from the Dataset tab or the Apify API.

Frequently Asked Questions

Do I need an account or login?

No. The Actor reads public video data and never logs in anywhere.

Which videos work?

Any public video that has captions - creator-provided or auto-generated. Shorts, youtu.be links, and bare IDs are all accepted.

What if a video has no captions?

The row is still written with metadata and transcriptStatus set to no-captions; the briefing fields are empty.

Can I get the transcript in another language?

Set language to the two-letter code you want. If that track does not exist, the video's default track is used and transcriptLanguage shows which one.

Is the full transcript included?

Yes. Every row contains the full transcript string and the transcriptSegments array with per-line timestamps.

Am I charged for failed results?

You are charged for video reports written to the dataset. Videos that fail before a report is written are not charged.

How fresh is the data?

Metadata and the transcript are read live at run time. viewCount is the value at read time.

Can it export to Google Sheets?

Yes. Connect the dataset to the Google Sheets integration or pull it from the API.

Does it handle duplicate videos?

Yes. Videos are keyed by ID; a duplicate entry is processed once.

Support

Questions and bug reports: use the Issues tab on the Actor's Console page. Include the input you ran and the run ID.

Legal / Responsible Use

You are responsible for complying with YouTube's Terms of Service, applicable law, and responsible data-use requirements. Transcripts and video metadata may include personal data; you are responsible for meeting any applicable privacy and data-protection obligations (for example GDPR) for personal data you collect and store, and for respecting the rights of content creators when reusing transcript text.