Twitch Video Search Scraper

Search public Twitch videos and export each result as an individual record. Twitch Video Search Scraper returns titles, channel identity, category, views, duration, dates and available preview images.

Find source-linked video candidates for editorial research or a catalog, then select URLs for detailed collection.

At a glance	Details
📥 Input	Enter one video search keyword or phrase per item. Each page returns up to 10 videos.
📤 Output	One video per success row, with source input retained
📄 Pagination	pagesPerSearch limits result pages per query. Each search request asks for 10 video results; unique output counts can be lower.
🔐 Login required	No
⚡ Processing	Up to 5 HTTP requests concurrently with automatic retries
💾 Delivery	One Apify dataset view, useful flat columns and complete nested JSON

What the Twitch Video Search scraper collects 🔎

The Actor follows Twitch video-search results for each query. Search results retain their original query context and video identity.

Data can include:

• Video ID, title and public URL
• Owner channel identity and category when supplied
• View count, duration and available dates
• Preview image references and video-type information when returned
• Original search input

How to collect video search results from Twitch 🚀

1. Enter one or more supported inputs in inputList.
2. Set the result-page budget for each input.
3. Start the Actor.
4. Open the Videos dataset view and review the individual records.
5. Export the dataset or retrieve it from your application.

{
  "inputList": [
    "minecraft"
  ],
  "pagesPerSearch": 1
}

pagesPerSearch limits result pages per query. Each search request asks for 10 video results; unique output counts can be lower.

Twitch video search results data and complete output 📦

Field	Type	Meaning
input	string	Original submitted input.
status	string	Result status: success or failed.
type	string or null	Normalized result entity type.
id	string or null	Identifier of this object, preserved as a string.
url	string or null	Public source URL for this object.
title	string or null	Public title supplied for this object.
created_at	string or null	Creation time in an ISO-formatted date string when supplied.
duration_seconds	integer or null	Duration in seconds, retaining fractional values when supplied.
view_count	integer or null	Visible view count at collection time; not a unique-viewer estimate.
preview_thumbnail_url	string or null	Available video preview-thumbnail URL.
preview_thumbnail_template_url	string or null	Source thumbnail URL template, including any size placeholders.
preview_blur_reason	string or null	Public reason given for blurring a preview, when supplied.
owner	object or null	Public channel that owns the selected video.
owner.id	string or null	Identifier of this object, preserved as a string.
owner.login	string or null	Twitch channel login used in its public URL.
owner.display_name	string or null	Public display name, preserving the source capitalization.
owner.url	string or null	Public source URL for this object.
owner.profile_image_url	string or null	Available public channel profile-image URL.
owner.primary_color_hex	string or null	Public channel theme color in hexadecimal form.
owner.is_partner	boolean or null	Public Twitch Partner flag when supplied.
game	object or null	Available game or category metadata for the parent object.
game.id	string or null	Identifier of this object, preserved as a string.
game.slug	string or null	Public identifier used in this object’s URL.
game.name	string or null	Public name of this blog or nested object.
game.display_name	string or null	Public display name, preserving the source capitalization.
game.url	string or null	Public source URL for this object.
game.box_art_url	string or null	Available category or game cover-image URL.

Every top-level output field appears in these complete examples. The success example is normalized from a public response; its values are a snapshot and can change. Nested arrays remain inside their parent result.

Complete representative success result:

{
  "input": "minecraft",
  "status": "success",
  "type": "video",
  "id": "2866178734",
  "url": "https://www.twitch.tv/videos/2866178734",
  "title": "MC Championship: Fall Stars - Central Stream!",
  "created_at": "2026-09-05T18:35:53Z",
  "duration_seconds": 12318,
  "view_count": 36734,
  "preview_thumbnail_url": "https://static-cdn.jtvnw.net/cf_vods/d3stzm2eumvgb4/2138acea6b4e8db7a664_minecraft_317675896420_1788633348//thumb/thumb0-214x120.jpg",
  "preview_thumbnail_template_url": "https://static-cdn.jtvnw.net/cf_vods/d3stzm2eumvgb4/2138acea6b4e8db7a664_minecraft_317675896420_1788633348//thumb/thumb0-{width}x{height}.jpg",
  "preview_blur_reason": "BLUR_NOT_REQUIRED",
  "owner": {
    "id": "112568845",
    "login": "minecraft",
    "display_name": "Minecraft",
    "url": "https://www.twitch.tv/minecraft",
    "profile_image_url": null,
    "primary_color_hex": null,
    "is_partner": true
  },
  "game": {
    "id": "27471",
    "slug": "minecraft",
    "name": "Minecraft",
    "display_name": "Minecraft",
    "url": "https://www.twitch.tv/directory/category/minecraft",
    "box_art_url": null
  }
}

Complete failed result:

{
  "input": "minecraft",
  "status": "failed",
  "type": null,
  "id": null,
  "url": null,
  "title": null,
  "created_at": null,
  "duration_seconds": null,
  "view_count": null,
  "preview_thumbnail_url": null,
  "preview_thumbnail_template_url": null,
  "preview_blur_reason": null,
  "owner": null,
  "game": null
}

A failed row preserves input, sets status to failed, and sets every other top-level field to null. The run log records the reason. Optional successful fields can be null or empty when Twitch does not supply them. Object fields are exposed through useful columns in the single Videos view; arrays are not expanded into extra result rows.

Input requirements and pagination settings ⚙️

Parameter	Type	Required	Default	Rules
inputList	array of strings	Yes	None	Enter one video search keyword or phrase per item. Each page returns up to 10 videos. At least 1 item.
pagesPerSearch	integer	No	1	Maximum result pages per input. Setup requests do not count as pages. Stops when the source has no next page or repeats a continuation. Actual rows per page can vary. Minimum 1.

Enter plain search queries, not URLs. This Actor searches video results; clips have a separate direct-details Actor.

Enter plain keyword phrases, such as gaming for channel search or minecraft for video search. Search URLs are not accepted. Each value is a separate search input, and only the result type of this Actor is collected.

Duplicate normalized inputs are processed once. Within the collection for one source input, repeated records are removed while distinct interactions or objects remain separate. The same object returned for different source inputs retains its source relationship.

The first result page counts as page 1. Public-page preparation and metadata lookups do not consume result pages. Collection stops at your page budget or the end of the available list. A short page alone does not mean that the list has ended.

Twitch video search results use cases 🎯

Video discovery

Find relevant video records from a topic query with source links for review.

Editorial video catalogs

Retain titles, channels, categories and dates for manual selection and indexing.

Metadata comparisons

Compare the returned duration, view and category fields in your own analysis.

Pricing and saved-result behavior 💰

Check the Actor’s Pricing tab for the active pricing model and current rate. Store settings can change, so this README does not claim a fixed price or runtime.

Under dataset-item pricing:

• Each successful result represents one video for its source input.
• An invalid or unavailable input, an input with no accessible results, or an exhausted request can produce a failed row.
• Retry attempts and supporting requests do not create extra dataset rows by themselves.
• Nested media, profile information and other arrays remain part of their parent row.
• A normal empty continuation after earlier successes creates no additional row.
• Saved failed rows are not assumed to be free; check their treatment in the active pricing configuration.

More inputs and larger page budgets can produce more saved rows. Start with a small run and check actual usage before increasing the workload.

Twitch video search results API 🔌

Replace $ACTOR_ID with the identifier shown in this Actor’s API tab and $APIFY_TOKEN with your Apify token.

curl -X POST "https://api.apify.com/v2/acts/$ACTOR_ID/runs?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"inputList":["minecraft"],"pagesPerSearch":1}'

Download the dataset in JSON, CSV, Excel or other formats available in the Console, or retrieve it through the Apify API. Use schedules, webhooks and Apify integrations to connect results to Google Sheets, Make, Zapier, cloud storage or your own application. These connections are configured separately by the user.

Reliability, retries, and public-data limits ⚠️

Collection uses pure HTTP with up to five concurrent requests. Invalid syntax and confirmed missing, removed or unavailable targets stop without unnecessary retries. Temporary network or proxy failures, timeouts, blocking responses, malformed data, throttling and server errors allow up to five total attempts per request.

Successful earlier pages remain saved if a later request fails. A normal end after saved results is not retried. An input with no accessible results can receive a failed row; that alone does not prove that its underlying source does not exist. An exhausted later request can add a failed row while preserving previous output.

Video search is not a complete channel archive or a clip-search promise. It returns metadata rather than video files, protected playback or VOD chat. Availability, ranking and counters can change.

Twitch controls public availability, ranking and optional fields. Result counts can be affected by duplicates, removed content and changes during a run. A page budget is a collection limit, not a promise of exhaustive coverage.

An invalid string within a valid input list does not stop other inputs. Invalid overall configuration, such as a non-string list item or a wrong page-count type, exits before source requests. Startup failures, unavailable dataset storage or an unrecoverable result-save error can still stop the whole run. A failed save is not retried as a new scraping request.

Frequently asked questions ❓

Does this search clips too?

No. The supported search result type is videos. Use Clip Details Scraper for known clip links.

Does it download videos?

No. It returns public video metadata and available preview references.

Does search return every video from a channel?

No. It follows the website’s query results within your page budget; it is not a channel archive.

Does it require a Twitch account or personal cookies?

No. You do not need to provide a Twitch account, password or personal session cookie. Collection uses HTTP without opening a browser.

What happens to invalid or unavailable inputs?

Invalid individual inputs are recorded as failed without a source request. Confirmed missing or unavailable targets stop without unnecessary retries. Temporary failures allow up to five total attempts per request; other inputs and earlier saved results remain available.

Can I export results or run the same input again?

Yes. Export the default dataset or retrieve it through the Apify API. You can create an Apify schedule and connect completed runs to your own workflow.

Support, responsible use, and related actors 🛟

For a reproducible issue, use the Actor’s Issues tab and provide the run ID, approximate time, a safe public input, expected behavior and actual result. Include the mode or page budget when relevant. Never share access tokens, personal cookies or proxy credentials.

Use public data responsibly and follow applicable privacy, copyright, contractual and platform requirements before storing, analyzing or redistributing collected information.