Twitch Channel, Clips, Videos & Stream Data — Try the interactive demo
A guided live test. Prefer the raw endpoints? See the API reference & playground →
Enter an API key from your Dashboard to enable the examples below. Requests send it to this platform for authentication.
Log in / Sign up · Dashboard →
Please paste your API key first.
Endpoints
Your API key is attached automatically to each call.
/api/project/twitch-scraper/list-channel-clips List Twitch Channel Clips by Login
Details
Returns the public Twitch clips for one channel as a cursor-paginated collection that is separate from the channel's videos and live-stream data. Inputs: channel (required) is the channel login shown in the public Twitch URL, for example ninja, not a display title; cursor (optional, number) is the continuation value returned by a previous response for the same channel and is omitted on the first request. Observed in a live probe for channel ninja: the response is the standard envelope {status, request_id, message, data} with status 'ok' and a null message, and data.user holds the channel id, a clips array of 12 entries and a top-level cursor (an integer in the probe). Each clips entry is {cursor, node}; the per-entry cursor was null for the first eleven entries and a string on the last one. Each node carried id, slug, url, embedURL, title, viewCount, language, thumbnailURL, createdAt, durationSeconds and isFeatured, plus nested curator (id, login, displayName), game (id, slug, name, boxArtURL), broadcast (id, startedAt, title) and broadcaster (id, login, displayName, profileImageURL, primaryColorHex, roles) objects; champBadge and guestStarParticipants were null. These field names come from a single probe and may not be present on every clip, and the next-page behavior of the cursor was not exercised in the probe. Clip availability changes over time: a result does not guarantee downloadable media or perpetual access, and the list should not be assumed to contain every historical clip.
| Parameter | In | Value |
|---|---|---|
| query | ||
| query |
Request cost is temporarily unavailable. No price is assumed.
/api/project/twitch-scraper/get-channel-details Get Twitch Channel Profile and Followers by Login
Details
Returns public profile metadata for one Twitch channel, looked up by its login. Use it to identify a channel before fetching clips, videos or stream information. Input: channel (required) is the channel login shown in the public Twitch URL, for example ninja, not a display title. Observed in a live probe for channel ninja: the response is the standard envelope {status, request_id, message, data} with status 'ok' and a null message, and data.user holds id, login, displayName, primaryColorHex, profileImageURL and bannerImageURL; followers.totalCount (an integer); roles with isPartner and isParticipatingDJ booleans; and a channel object with its own id, self (isAuthorized, restrictionType), trailer (video) and home.preferences.heroPreset. The stream field was null in the probe because the channel was offline at that moment; the populated live-stream branch was not observed, so its field names are not documented here and live status should be read from the stream operations. Identifiers should be kept as strings and optional profile fields kept as returned: a missing field does not prove the channel was deleted, and a channel that exists is not necessarily broadcasting.
| Parameter | In | Value |
|---|---|---|
| query |
Request cost is temporarily unavailable. No price is assumed.
/api/project/twitch-scraper/search-channels Search Twitch Channels by Name or Keyword
Details
Searches publicly discoverable Twitch channels by text and returns a short list of candidates so you can pick the intended channel. Input: query (required) is a channel name or free search text, for example ninja. No cursor or page parameter is exposed, and the response carries no cursor. Observed in a live probe for query ninja: the response is the standard envelope {status, request_id, message, data} with status 'ok' and a null message, and data.channels held 15 entries, each {trackingID, item}. Each item carried id, login, displayName, profileImageURL, description (a string, or null on 2 of the 15), followers.totalCount, roles (isPartner, isParticipatingDJ), broadcastSettings (id, title), lastBroadcast (id, startedAt), channel (id, schedule), latestVideo.edges[].node (id, title, lengthSeconds, preview thumbnail fields) and topClip.edges[].node (id, slug, title, durationSeconds, thumbnailURL); self, watchParty and channel.schedule were null. item.stream was null for one offline channel and an object for the other 14 in this search probe, with fields such as type, viewersCount, game and freeformTags; those live-stream fields were seen only in this search sample and are not verified for other operations. Keep the returned login and source ids: a display name may differ from the login that channel and stream lookups require. Results can change between calls and are not an exhaustive directory.
| Parameter | In | Value |
|---|---|---|
| query |
Request cost is temporarily unavailable. No price is assumed.
/api/project/twitch-scraper/get-stream-details Get Twitch Stream Details by Channel Login
Details
Looks up the current broadcast of one Twitch channel and returns it as a point-in-time snapshot. Input: channel (required) is the channel login shown in the public Twitch URL, for example ninja, not a display title. Observed in a live probe for channel ninja while the channel was offline: HTTP 200 with the standard envelope {status, request_id, message, data}; status was 'ok', message carried a notice that the channel is currently offline, and data.user held id (a string), stream set to null and __typename 'User'. A null stream therefore means the channel is not broadcasting, or that no stream payload was available at request time; it is a valid result, not a failed request, so use status and message to tell the two apart. The populated live-stream branch was not observed in the probe, so its field names are not documented here and should be checked against a real live response before you map them. The operation does not record or download the stream and does not guarantee continuous live status; for the audience figure use Get Stream Viewer Count, and for classification labels use List Stream Tags.
| Parameter | In | Value |
|---|---|---|
| query |
Request cost is temporarily unavailable. No price is assumed.
/api/project/twitch-scraper/list-stream-tags List Twitch Stream Tags by Channel Login
Details
Returns the classification tags attached to the current stream of one Twitch channel, so you can group or filter broadcasts by the labels Twitch shows. Input: channel (required) is the channel login shown in the public Twitch URL, for example ninja, not a display title. Observed in a live probe for channel ninja while the channel was offline: HTTP 200 with the standard envelope {status, request_id, message, data}; status was 'ok', message was null, and data.user held id (a string), stream set to null and __typename 'User'. With a null stream there were no tags to read, and that is a documented valid result rather than an invalid channel. The populated branch was not observed in the probe, so the names of the tag identifier and label fields are not documented here; inspect a response from a live channel before mapping them, keep the returned identifiers and labels exactly as supplied and never infer tags that are not returned. Tags can change over time and do not prove the channel is still live when you display them later; use Get Stream Details for the broadcast context.
| Parameter | In | Value |
|---|---|---|
| query |
Request cost is temporarily unavailable. No price is assumed.
/api/project/twitch-scraper/get-stream-viewer-count Get Twitch Stream Viewer Count by Channel Login
Details
Returns the viewer-count information available for the current stream of one Twitch channel, as a snapshot taken at request time. Input: channel (required) is the channel login shown in the public Twitch URL, for example ninja, not a display title. Observed in a live probe for channel ninja while the channel was offline: HTTP 200 with the standard envelope {status, request_id, message, data}; status was 'ok', message was null, and data.user held id (a string), stream set to null and __typename 'User'. With a null stream there was no viewer figure in the response, which is a documented valid result. The populated live-stream branch was not observed in the probe, so the name of the viewer-count field is not documented here; inspect a response from a live channel before mapping it, and keep a missing value distinct from numeric zero. The number is the audience of the current stream, which is different from historical video views or total channel followers. The operation provides no audience history or time series, so record your own retrieval timestamp with every snapshot.
| Parameter | In | Value |
|---|---|---|
| query |
Request cost is temporarily unavailable. No price is assumed.
/api/project/twitch-scraper/list-channel-videos List a Twitch channel's public past-broadcast videos
Details
Returns the public video (past broadcast) collection that is currently available for one Twitch channel, looked up by channel login. Send the required channel parameter (for example ninja) and, on later pages only, the optional numeric cursor returned for the same channel. The response is wrapped as {status, request_id, message, data}, and the records sit under data.user. Observed in a live probe for channel ninja: data.user.id (channel id), data.user.videos (12 edges per page), and data.user.cursor (an integer, 2 on the first page) as the page-level continuation value. Each edge carries an opaque cursor string and a node with id, title, publishedAt (ISO 8601 UTC), lengthSeconds, viewCount, previewThumbnailURL, animatedPreviewURL, game (id, name, displayName, slug, boxArtURL), owner (id, login, displayName, profileImageURL, primaryColorHex, roles.isPartner), self.isRestricted, resourceRestriction and contentTags. In that probe the videos came back newest first, resourceRestriction was null and contentTags was an empty list; populated values for those two fields, the second page, and the end of the collection were not observed. Pagination: omit cursor on the first request, keep channel identical, reuse the returned continuation value unchanged, stop when it is missing or repeats, and deduplicate by video id. Scope: the response reflects available video records, not a complete historical archive; deleted, expired or restricted recordings may be absent. The operation returns metadata and image links only and does not download media.
| Parameter | In | Value |
|---|---|---|
| query | ||
| query |
Request cost is temporarily unavailable. No price is assumed.