YouTube Channel Scraper — Videos, Shorts & Streams

Paste YouTube channel URLs, handles or channel IDs to export public videos, Shorts and streams as structured JSON, CSV or Excel—without a YouTube login or YouTube API key. Pay $0.40 per 1,000 successfully delivered video rows, including platform usage; tier discounts are listed below. Only valid unique video rows successfully written to the Dataset generate the result event.

• Choose the content you need: regular videos, Shorts and streams, with per-channel limits and native latest, popular or oldest ordering when the category supports it.
• Keep useful context: video titles, links, duration, original view-count text and available channel identity stay together.
• Add channel profiles: available descriptions, public links, images and subscriber metadata are saved separately from video rows.
• Control extra detail work: full descriptions are optional; a failed description lookup preserves the valid listing row and reports incomplete detail.
• Avoid duplicate results: each video ID is delivered once per run, including overlapping channel aliases.

Quick start

1. Enter a public channel URL you already know.
2. Keep a small result limit for the first run.
3. Open the Dataset for video rows, CHANNELS.json for channel profiles and OUTPUT for status.

{
  "channels": ["https://www.youtube.com/@NASA"],
  "tabs": ["videos"],
  "sort": "latest",
  "maxResults": 3,
  "includeChannelInfo": true,
  "includeDescription": false
}

Empty API input succeeds with zero output rows and zero result charges. Empty API input does not collect a sample automatically. The Console form prefills NASA for an intentional test run. A new run starts a new Dataset; it is not a continuation of an earlier run.

What you receive

The Dataset keeps the native v1 video contract. Missing public evidence stays null rather than becoming an invented zero, date or verification flag.

Output	Useful fields and meaning
Video identity	id, canonical url, title, contentType, sourceChannel
Channel context	channelId, channelUrl, channelTitle
Popularity	views, original viewsText, and viewsPrecision; abbreviations are not exact counts
Date evidence	publishedText, displayedDate, dateSource; publishedAt is null because exact timestamps are not inferred
Duration and image	durationSeconds, durationText, thumbnail, where available
Details and ordering	description, descriptionComplete, detailsStatus, sortRequested, sortStatus, liveStatus
Channel profile record	CHANNELS.json: available public description, links, subscribers, counts, images and status
Run summary	OUTPUT: delivery, errors, requests, date-filter uncertainty and budget status

Real selected output

The following selected fields come from the NASA input above, captured on October 13, 2026 from accepted private build 0.1.3. It is a historical sample, not a forecast of the next run's titles or counts. Descriptions were not requested; approximate listing counts were preserved as text.

{
  "contractVersion": 1,
  "id": "ErAqN6gXqZQ",
  "url": "https://www.youtube.com/watch?v=ErAqN6gXqZQ",
  "title": "NASA's SpaceX Crew-13 Mission (Official NASA Trailer)",
  "contentType": "video",
  "channelTitle": "NASA",
  "description": null,
  "descriptionComplete": false,
  "publishedText": "12d ago",
  "publishedAt": null,
  "views": null,
  "viewsText": "101K",
  "viewsPrecision": "abbreviated_or_unparsed",
  "durationSeconds": 62,
  "durationText": "1:02",
  "detailsStatus": "not_requested",
  "sortRequested": "latest",
  "sortStatus": "confirmed",
  "liveStatus": null
}

Thumbnail and profile image links may change or expire; do not treat them as permanent assets. This Actor does not download videos, audio, images or screenshots.

Input reference

Input	Default	How to use it
channels	Empty	Public handles, IDs or channel URLs. Exact repeated strings are removed.
tabs	["videos"]	videos, shorts, streams, processed in supplied order under a shared cap.
sort	latest	latest, popular, oldest within each category, where native ordering is exposed.
maxResults	30	Limit per distinct trimmed input reference, 1–1000 in legacy mode. This is a cap, not a promised row count.
includeDescription	false	Request full public descriptions for selected unique rows. Unavailable detail retains the base row.
includeChannelInfo	false	Request extended profile fields, separately saved in CHANNELS.json.
maxResultsShorts / maxResultStreams	Unset	Supplying either activates independent limits: videos=maxResults, Shorts/streams default to 0. Each category accepts 0–1000.
oldestPostDate	Unset	YYYY-MM-DD at UTC midnight, or N day / N days relative to run start. Forces latest ordering.
outputFormat	v1	streamers adds a separate COMPATIBILITY.json record; it does not replace or change the Dataset.
startUrls	Unset	Migration alternative to channels; public channel URLs only. Do not supply both.
sortVideosBy	Unset	Migration alternative to sort: NEWEST, POPULAR, OLDEST. Do not supply both.

For profiles without video rows, use all three independent limits as zero:

{"channels":["https://www.youtube.com/@NASA"],"maxResults":0,"maxResultsShorts":0,"maxResultStreams":0,"includeChannelInfo":true}

Different references such as a channel URL and its @handle keep separate input quotas. Their shared video IDs are skipped globally; a later distinct video can fill the second quota. Only identical trimmed strings are merged.

maxResults:0 alone remains invalid. Independent limits override tabs and process videos, then Shorts, then streams. Categories are not merged into one chronological feed. The advanced detailConcurrency control defaults to 8, accepts 1–32 and preserves listing order.

Pricing

The result-delivered event is charged only after a valid unique video row is successfully written to the Dataset. These prices include platform usage:

Apify tier	USD per 1,000 successful video rows
FREE	$0.40
BRONZE	$0.38
SILVER	$0.36
GOLD	$0.34
PLATINUM	$0.34
DIAMOND	$0.34

Channel profiles included with video rows have no additional fee. Regular videos, Shorts and streams use the same result-delivered event, only after a valid unique row is successfully committed to the Dataset. Empty input, invalid inputs, unavailable or failed listings and duplicates do not trigger that event. No start fee or minimum charge is part of this decision. At the approved FREE rate, three delivered rows would cost $0.0012; 1,000 would cost $0.40. These are price illustrations, not actual run costs or charged-run evidence.

Channel-only collection writes profile records to CHANNELS.json and no video rows, so it generates no result-delivered event. Full-description enrichment uses the same delivered-video event and has no separate description fee. A valid listing row is still delivered and charged when its optional description lookup is unavailable; inspect descriptionComplete, detailsStatus and OUTPUT for the outcome.

Reliability and limitations

On October 14, 2026, build 0.1.8 delivered 1,000 ordinary NASA rows in each of two runs and 100 complete-description rows in each of two smaller runs. Shorts returned 52 and 53 of requested 100 with budget-exhaustion errors despite platform success. Check OUTPUT.stoppedOnBudget, errors and delivered counts; a successful platform status alone does not establish complete collection. An abort attempted after 30 observed rows arrived after the run completed 60; interrupted recovery was not verified, and no replacement or resurrection was run. These measured outcomes do not establish universal completeness or stability.

On October 13, 2026, build 0.1.3 delivered three unique NASA video rows with channel profiles, and the separate compatible export attached profile metadata to all three. A prior private run failed a Dataset enum check; the schema was corrected and the second run passed. This small smoke does not establish universal coverage, stability or a speed advantage.

• Public, unavailable, restricted or removed content may return fewer rows or errors. Other valid channels can continue.
• Unknown dates and relative-age intervals crossing a cutoff are retained and reported as incomplete date filtering. Month/year labels do not become exact timestamps; a complete date range is not guaranteed.
• sortStatus distinguishes confirmed native selection from requested ordering without returned selection evidence. Unsupported sorts are reported rather than silently replaced.
• liveStatus is live, upcoming or null; live audience and accumulated view counts are different measures.
• Counts may be abbreviated, fields absent and metadata partial. Compatibility field names do not mean every competitor field or value is identical.
• Interrupted or budget-limited collection retains a resumable state when available. The same run can reuse its Dataset and skip delivered IDs, but the October 14 live test did not verify an actual aborted-run resurrection. New runs have independent output and billing.
• Failed Dataset writes or state persistence stop the run. A charge failure is reported and is not retried for an already delivered row.

Automation and integrations

Use Apify's Actor API to start runs, download Dataset exports and inspect run status. Add schedules for periodic monitoring and webhooks for downstream processing; connect results to spreadsheets, databases or supported Apify integrations. Keep a small limit when testing an automated workflow. This is a batch Actor, not a live web server.

See the official Actor API documentation and Dataset export documentation.

FAQ

Do I need a YouTube login or API key? No. This collects available public metadata, not private account content.

Why is the Dataset empty for profile-only input? Profiles are in CHANNELS.json; they do not become video rows or result events.

Why is an exact view count or publication timestamp missing? Listing abbreviations and relative ages are preserved honestly. publishedAt stays null; available watch calendar dates are kept separately.

Can I migrate an existing channel-URL input? Use startUrls instead of channels, sortVideosBy instead of sort, and optionally outputFormat:"streamers". Read the compatibility wrapper's items array; the native Dataset remains unchanged.

Does an empty result mean the whole channel has no videos? No. Inspect OUTPUT for category errors, partial progress, limits and date uncertainty before drawing that conclusion.

Can I continue a stopped run? When resurrection is supported, use that same run so its Dataset and saved state remain available. Actual aborted-run recovery was not verified in the October 14 test. Starting a new run does not share prior delivery IDs.

Responsible use and independence

Use public metadata responsibly and comply with applicable rules and YouTube's terms. This independently developed Actor is not affiliated with, endorsed by or operated by YouTube or Google. Platform names and logos identify the supported public data source.

Streamers compatibility format 2

outputFormat: "streamers" adds COMPATIBILITY.json with formatVersion: 2 and datasetContractVersion: 1. Read migration rows from items: regular videos have 31 fields, Shorts 30 (without duration), and scheduled streams 32. Channel profiles and aboutChannelInfo have 16 fields. Missing profile booleans default to false in this projection; their original unknown values remain null in diagnostics. Native Dataset 25 fields, OUTPUT and CHANNELS retain source evidence.

For existing format 1 readers, move items[index].compatibility.* to diagnostics.items[index].*, including count precision, concurrent viewers, date sources and sort status. Video descriptions/completion and live/ended stream states also move to diagnostics; scheduled streamType remains in items. Channel diagnostics move from channels[index].compatibility.* to diagnostics.channels[index].*. Diagnostic index and channelIndex identify associations; missing or conflicting profiles are never attached by guesswork. Public progressKey and standardizedUrl derive from the source channel about URL, with no copied session identifiers.