§ 01
Audience & creator discovery
Build lead and creator-discovery pipelines from Twitch public profiles. Filter by follower count, engagement, and region. The JSON is flat enough to drop into a Postgres column without normalisation.
§ The Twitch hub
Twitch
scraper.
No proxy needed. 4 pre-built endpoints, parsed JSON in seconds.
4 endpoints·From $0.003 per call·Median under 2s
§ One call
Send a handle.
Get back JSON.
The profile scraper takes one input: handle. The response carries every public field Twitch exposes.
Request
cURL
curl https://api.hproxy.com/v1/scrape/twitch/profile/by/handle \
-H "Authorization: Bearer $HPROXY_KEY" \
-H "Content-Type: application/json" \
-d '{"handle":"ishowspeed"}'POST // 1 call · $0.0030 · <2s
Response200 OK
{
"success": true,
"id": "220476955",
"isLive": true,
"currentViewersCount": 6293,
"handle": "ishowspeed",
"displayName": "IShowSpeed",
"profileImageURL": "https://static-cdn.jtvnw.net/jtv_user_pictures/….png",
"description": "YouTube:IShowSpeed Instagram:IShowSpeed: Snapchat:wishowspeed",
"followers": 3830874,
"isPartner": true
}What comes backparsed fields
Twitch Profile
twitch_profile
idstring
isLiveboolean
currentViewersCountnumber
handlestring
displayNamestring
profileImageURLurl
descriptionstring
One real call · one structured record
§ Coverage
What you can
actually pull.
§ 01
Profiles
Bio, follower and following counts, post totals, verified flag, links, and the profile image — every public field the platform exposes about an account.
§ 02
Videos
Video metadata: title, description, view and like counts, duration, channel reference, publish date, and the thumbnail URL.
§ 03
Clips
Short clip objects with title, creator, view count, and the playback URL.
§ 04
Users
Public user objects with handle, display name, follower counts, and the profile image.
§ Every endpoint
The full list.
Prices included.
Every line is a live endpoint. Click through for the input schema, response shape, and the curl recipe.
Twitch Profile
/v1/scrape/twitch/profile/by/handle
Public Twitch user profile with stream + clips.
$0.0030
per call
Twitch User Videos
/v1/scrape/twitch/user/videos/by/handle
VOD list for a Twitch user with optional filter and sort.
$0.0030
per call
Twitch Clip
/v1/scrape/twitch/clip/by/url
Detailed metadata for a Twitch clip with download URL.
$0.0030
per call
Twitch User Schedule
/v1/scrape/twitch/user/schedule/by/handle
Upcoming streaming schedule for a Twitch user.
$0.0030
per call
§ What people build
One Twitch scraper.
Many shapes of work.
§ 02
Brand & trend monitoring
Track every public post mentioning a brand or keyword across Twitch. Pipe captions into sentiment models and chart engagement over time without writing a single parser.
§ 03
AI training data
Feed clean, structured Twitch JSON straight into your model fine-tuning pipeline. Same schema every call — no scraping infrastructure to maintain.
§ 04
Dashboards & reporting
Wire Twitch signals into internal dashboards. One API key, predictable JSON, dollar-denominated pricing — finance can reason about the bill.
§ Questions
The honest
answers.
The questions that come up before the first call. If yours isn’t here, the founder reads support email himself — just write in.
Do I need proxies to use the Twitch scraper?
No. Every endpoint runs on our residential pool — fifty million IPs sourced through opt-in partner SDKs. You send an HTTP request with your API key; rotation, retries, and anti-bot handling are ours to worry about.
How does HProxy get past Twitch’s rate limits and bot detection?
Each call is routed through a residential session that matches typical organic traffic patterns. We don’t advertise specifics; what matters is the result — a 99%+ success rate on public-data endpoints in real customer traffic.
What does the JSON response look like?
Every endpoint returns a flat top-level wrapper (platform, scraper, data, creditsUsed, elapsedMs, requestId) with the parsed Twitch entity inside data. See the live example near the top of this page for the exact shape.
Is scraping public Twitch data legal?
Public data scraping (no login, no private content, no PII beyond what twitch.tv itself publishes) sits within the boundaries set by hiQ Labs v. LinkedIn and similar US/EU precedents. We don’t serve endpoints that touch authenticated content. You’re responsible for your own use of the data under your local data-protection law (GDPR, CCPA, etc.).
What does each call cost?
From $0.003 per call on the standard endpoints. Endpoints that do heavier upstream work — AI transcripts, posts with media archived to R2 — cost more. The exact per-call price shows on each endpoint row below. One call = one charge, never multiplied by record count.
Can I test before signing up?
Yes — the wallet starts with a $2 deposit that’s refundable for 24 hours. That’s plenty of calls to verify the response shape matches what you’re building across all 4 Twitch endpoints.
How fast are the responses?
Median response time across the Twitch endpoints is under 2 seconds. Heavier endpoints (deep comment threads, cursor-paginated lists) can run 4-8 seconds depending on the page depth requested.
§ Try it
Two dollars,
first run.
Sign in with Google, drop $2 in the wallet, fire your first Twitch call against twitch.tv. If it doesn’t do what you need, the balance is refundable for 24 hours.