YouTube Keyword Monitor — New Videos Only | $1.99 / 1K videos | No API Key

Watch YouTube for new videos on your keywords, or from the channels you follow — and get only what appeared since your last run.

No API key. No quota. No login. No cookies. This Actor reads public YouTube pages, so there is no Google Cloud project to set up, no daily quota to run out of, and no account to get flagged. If you have hit the YouTube Data API's daily unit limit, this is a drop-in alternative that does not have one.

$1.99 per 1,000 videos returned. Platform usage included — no second bill for compute or proxy.

---

This is a monitor, not a bulk scraper

Most YouTube scrapers are built to dump everything: give them a channel, get 20,000 videos back. That is a different job.

This one is built to run on a schedule and answer one question: what is new?

	This Actor	A bulk scraper
Built for	Running every 15 min, forever	One big extraction
Returns	Only videos you have not seen	Everything, every time
Repeat cost	You pay for new videos only	You pay for the same videos again
Dedupe	Built in, 7-day memory	You do it yourself
Typical run	A handful of new videos	Thousands of rows

If you want a one-off bulk dump, use a bulk scraper. If you want an alerting feed, this is it.

---

What you can do with it

• Brand and competitor monitoring — know within minutes when someone uploads a video mentioning your product
• Trend watching — track a topic and collect every new video about it over weeks
• Creator tracking — follow a list of channels and get their uploads as they land
• Lead and partner discovery — find the channels publishing about your niche
• AI agents and RAG — a clean, deduplicated feed of fresh video metadata
• Alerting — pipe new videos into Slack, Discord, email or a webhook

---

Two inputs, mix freely

Keywords

{
  "keywords": ["ai agents", "web scraping"],
  "sort": "date",
  "onlyNew": true
}

sort: date puts newest uploads first, which is what monitoring needs. relevance, views and rating are also available when you want ranking instead of recency.

Channels

{
  "channels": ["@NASA", "@veritasium"],
  "onlyNew": true
}

Handles work with or without the @, and full channel URLs are accepted too.

Both at once

{
  "keywords": ["mars rover"],
  "channels": ["@NASA"],
  "sort": "date",
  "onlyNew": true,
  "maxAgeHours": 48
}

Every item is tagged with _source (search or channel) and _query so you can split them downstream.

---

Ready-made examples

Brand monitoring, every 15 minutes

{
  "keywords": ["acme corp", "acme review"],
  "sort": "date",
  "onlyNew": true,
  "maxAgeHours": 24
}

Follow a set of creators

{
  "channels": ["@mkbhd", "@LinusTechTips", "@veritasium"],
  "onlyNew": true
}

Only videos that already have traction

{
  "keywords": ["product launch 2026"],
  "sort": "views",
  "minViews": 10000,
  "onlyNew": true
}

Full snapshot, no dedupe (one-off research)

{
  "keywords": ["electric vehicle review"],
  "sort": "relevance",
  "onlyNew": false
}

Feed an AI agent

{
  "keywords": ["the topic your agent is asking about"],
  "sort": "date",
  "onlyNew": false,
  "maxAgeHours": 24
}

---

Input reference

Field	Type	Default	What it does
keywords	array	—	Search terms to monitor.
channels	array	—	Channel handles or URLs to monitor.
sort	string	date	date (newest, for monitoring), relevance, views, rating.
onlyNew	boolean	true	Return only videos not seen in previous runs. 7-day memory per input set.
maxAgeHours	integer	0	Skip videos older than this. 0 = no limit. Approximate — see below.
minViews	integer	0	Skip videos below this view count.
maxResults	integer	0	Videos per keyword / channel. 0 = first page only. Higher values page through YouTube.
includeShorts	boolean	true	Shorts are often more than half of a keyword's results. Turn off for long-form only.
filterLanguages	array	[]	Keep only titles written in these languages. Undetermined titles are kept.
enrichVideos	boolean	false	Open each video for its exact publish time, like count, full description and keywords. One extra request per video.
transcripts	boolean	false	Add the spoken transcript of each video.
transcriptLanguages	array	["en"]	Preferred transcript languages, tried in order.
proxyConfiguration	object	Apify residential	Residential recommended.
concurrency	integer	3	How many keywords/channels to fetch at once.

---

Output

One dataset item per video. Export as JSON, CSV, Excel or XML, or pull it straight from the Apify API.

Field	Type	Notes
video_id	string	YouTube video ID
url	string	Direct watch link
title	string
channel	string	Channel name
channel_url	string or null	Filled for both keyword and channel runs
duration_text	string or null	e.g. 14:05. Blank for a live stream.
view_count	integer or null	Parsed to a number
view_count_text	string or null	As YouTube shows it, e.g. 486K views
published_text	string or null	As YouTube shows it, e.g. 12 days ago
published_ts_approx	integer or null	Approximate Unix timestamp — see below
description_snippet	string or null	Present for keyword search only — YouTube does not render descriptions on a channel's video grid
thumbnail	string or null	Highest-resolution thumbnail URL
is_short	boolean	Whether this is a YouTube Short
language	string or null	Detected language of the title, ISO 639-1. null when it cannot be determined.
published_at	string or null	Exact ISO 8601 publish time. Only with enrichVideos.
published_ts	integer or null	Exact Unix timestamp. Only with enrichVideos.
like_count	integer or null	Only with enrichVideos.
description	string or null	Full description. Only with enrichVideos.
keywords	array	Video tags. Only with enrichVideos.
channel_id	string or null	Only with enrichVideos.
subscriber_count	integer or null	The channel's subscriber count, read from the video owner block. YouTube omits that block on some page variants, so this is null more often than the other fields — it is left empty rather than filled from a recommended channel. Only with enrichVideos.
channel_thumbnail	string or null	The channel's avatar.
channel_username	string or null	The channel's @handle.
comment_count	integer or null	Exact count. Only with exactCommentCount.
comment_count_approx	integer or null	Rounded — YouTube only shows 4.2K on the video page. Only with enrichVideos.
hashtags	array	Hashtags parsed from the description. Only with enrichVideos.
description_links	array	Links parsed from the description. Only with enrichVideos.
duration_seconds	integer or null	Only with enrichVideos.
transcript	string or null	Full spoken transcript. Only with transcripts.
transcript_status	string	Why a transcript is missing: ok, disabled_by_uploader, no_transcript, video_unavailable, blocked, error.
transcript_segments	array	Transcript with start, duration and text per line. Only with transcripts.
transcript_language	string or null	Language actually returned.
transcript_is_generated	boolean	Whether YouTube auto-generated it, as opposed to a human-written one.
_source	string	search or channel
_query	string	The keyword or channel that produced this item
_mode	string	The sort mode, or videos for channels

Sample record

{
  "video_id": "IwZVXmQdX1E",
  "url": "https://www.youtube.com/watch?v=IwZVXmQdX1E",
  "title": "NASA Moon Base: The First Six Months",
  "channel": "NASA",
  "channel_url": "https://www.youtube.com/@NASA",
  "duration_text": "14:05",
  "view_count": 486000,
  "view_count_text": "486K views",
  "published_text": "12 days ago",
  "published_ts_approx": 1757600000,
  "thumbnail": "https://i.ytimg.com/vi/IwZVXmQdX1E/hq720.jpg",
  "_source": "channel",
  "_query": "NASA",
  "_mode": "videos"
}

---

Turn it into a live monitor

1. Run once with your keywords and check the output.
2. Save it as a Task.
3. Add a Schedule — every 15 minutes works well.
4. Keep onlyNew on, so each run emits only videos you have not seen.
5. Connect the dataset to Slack, Discord, a webhook, Google Sheets, Make or Zapier.

You now have a YouTube alerting feed that never repeats itself.

Use it from an AI agent

Callable from the Apify API and over MCP. Set onlyNew to false for agent calls so each question gets a full answer rather than a delta.

---

What to expect

About timestamps — read this one. A search or channel listing shows only relative text like 12 days ago, never an exact time. The video's own page does carry an exact timestamp, and enrichVideos fetches it — so if you need real timestamps, turn that on. This Actor gives you that text verbatim in published_text, and converts it to an approximate Unix timestamp in published_ts_approx. The field is named approx on purpose. It is accurate enough to answer "in the last 24 hours" and not accurate enough to sort videos uploaded minutes apart.

Shorts carry no publish time at all. YouTube's Shorts component does not include one, so published_text, published_ts_approx and duration_text are always null for a Short. Since videos without a timestamp are kept rather than dropped, maxAgeHours does not constrain Shorts — set includeShorts to false if you need the age filter to apply to everything you receive.

Without enrichVideos, maxAgeHours is therefore an approximate filter. Videos with no readable publish time are kept, not dropped — you should not lose a video because YouTube was vague about its age.

How many results per run. With maxResults at 0, a keyword search returns roughly 15 to 50 videos depending on how many Shorts YouTube mixes in, and a channel returns its latest ~30 uploads. Raise maxResults to go deeper. With onlyNew on, a scheduled run typically returns far fewer — which is the point.

Why a run can return few or zero items. In order of likelihood:

1. onlyNew is on and nothing new has been uploaded. This is correct behaviour, not a failure.
2. maxAgeHours or minViews filtered everything out.
3. The keyword is genuinely quiet.
4. The channel handle is wrong or the channel has no public uploads.

Public content only. Private, unlisted and members-only videos are not accessible.

---

FAQ

Do I need a YouTube Data API key? No. That is the main reason to use this. No Google Cloud project, no quota units, no daily limit to exhaust.

Do I need to log in or supply cookies? No. The Actor reads public pages only.

Can my Google account get banned? There is no account involved, so there is nothing to ban.

How is this different from a YouTube scraper? A scraper dumps everything every time. This one remembers what it already gave you and returns only new videos. That makes it cheap to run every 15 minutes, and it means you are not charged twice for the same video.

Am I charged for videos I already received? No. With onlyNew on, videos you have seen are filtered before they reach the dataset, so they are never charged.

My second run returned almost nothing. Is it broken? Almost certainly not — that is onlyNew working. Set it to false if you want the full list every time.

How far back does the memory go? 7 days per input set. Change the keywords or channels and it starts a fresh memory.

How many videos actually have a transcript? Across 172 videos from four keywords, about 60% did. The rest is not a failure on this Actor's side: uploaders can switch captions off, and most Shorts have them off. Every video carries a transcript_status saying which it was, so an empty transcript is never ambiguous. Turn includeShorts off and the share rises sharply.

Does it get transcripts? Yes — set transcripts to true. If a video has no transcript in your preferred languages, whatever transcript it does have is returned rather than nothing, and a human-written one is preferred over an auto-generated one where there is a choice.

Does it get comments? Not in this version.

Does it include Shorts? Yes, and they are flagged with is_short so you can split or drop them. For many keywords Shorts are more than half of what YouTube returns — a search for skincare came back with 38 Shorts and 8 long-form videos. Set includeShorts to false for long-form only.

Can I filter by date? Use maxAgeHours. Read the timestamp note in What to expect first — it is approximate by necessity.

Which proxy should I use? Residential, which is the default. YouTube serves a stripped page to flagged IPs rather than returning an error, so a clean IP matters.

Can I use my own proxies? Yes, through proxyConfiguration.

---

Changelog

• 0.4.0 — Transcripts. Full text and timed segments, with a language fallback so a video is not skipped just because it has no English transcript.
• 0.3.0 — enrichVideos adds the exact publish timestamp, subscriber count, rounded comment count, hashtags, description links, like count, full description, keywords and channel ID by opening each video. The note claiming YouTube never exposes an exact upload time was wrong and has been corrected.
• 0.2.0 — Shorts are now returned and flagged with is_short (they were silently dropped before, which cost a keyword search most of its results). maxResults pages past the first page. Every video carries a detected language, and filterLanguages keeps only the languages you want.
• 0.1.1 — Channel runs now return duration_text, and view counts and publish times survive YouTube's compact metadata layout (previously null on roughly half of all channel runs).
• 0.1.0 — First release. Keyword and channel monitoring, cross-run dedupe, age and view filters.

---

Notes and limits

• Reads public YouTube pages only.
• By default a keyword search returns roughly the first page and a channel its latest uploads. Raise maxResults to page further — a channel's back catalogue keeps going for hundreds of videos.
• YouTube changes its page structure regularly. This Actor parses several of YouTube's video components and is maintained against those changes. Run a small sample first, check the output, then scale up.
• Timestamps are approximate. See What to expect.

---

Our other Actors

• Threads Scraper — search Threads by keyword or pull a profile's posts, replies, reposts and media. No login, no cookies.
• YouTube Transcript Scraper — the spoken transcript of any video, by URL, channel or keyword. Falls back to any available language instead of failing.

Disclaimer

This Actor collects only publicly available data. You are responsible for how you use the data you collect, including compliance with YouTube's terms of service, applicable data protection law such as GDPR, and any restrictions that apply in your jurisdiction. If you are unsure, take legal advice before scraping personal data at scale.

---

Keywords: youtube scraper, youtube monitor, youtube data api alternative, youtube without api key, no quota, youtube keyword alerts, youtube new video alerts, youtube channel scraper, youtube search scraper, extract youtube videos, export to csv, export to excel, json output, video monitoring, social listening, brand monitoring, youtube data extraction.