Twitch VOD Chat Scraper — Full Chat Replay Export avatar

Twitch VOD Chat Scraper — Full Chat Replay Export

Pricing

from $0.10 / 1,000 results

Go to Apify Store
Twitch VOD Chat Scraper — Full Chat Replay Export

Twitch VOD Chat Scraper — Full Chat Replay Export

Export the complete chat replay of any Twitch VOD — every message with its timestamp, username, badges and emotes. Analyze sentiment, find hype moments, and study communities. Just paste a VOD URL, no code or token needed.

Pricing

from $0.10 / 1,000 results

Rating

0.0

(0)

Developer

Haketa

Haketa

Maintained by Community

Actor stats

0

Bookmarked

14

Total users

2

Monthly active users

12 days ago

Last modified

Features

Limited permissions

Share

Twitch VOD Chat Scraper — Export Full Chat Replay (Timestamped Messages, Users, Emotes & Badges)

Apify Actor

This Twitch VOD chat scraper exports the complete chat replay of any Twitch VOD — every message with its exact stream timestamp, offset in seconds, username, stable user ID, subscriber/moderator/VIP badges, emotes, emote IDs, and chat color. Paste one or more VOD URLs or enter channel names, and this Twitch scraper walks the past broadcast from start to finish and returns clean, deduplicated rows ready for spreadsheets, analytics pipelines, or AI datasets. No login, no OAuth token, and no coding required.

Export your Twitch chat data directly to JSON, CSV, Excel, XML, RSS, or via the Apify API.

💬 Built for clippers and highlight editors, streamers and stream analysts, esports organizations, talent & influencer marketing agencies, academic researchers, and AI/ML engineers who need structured Twitch VOD chat at scale.


⚡ At a glance

FeatureDetails
🎯 Target sourcePublic past broadcasts & highlights on Twitch.tv
🔍 Input modesDirect VOD URLs / Video IDs (videoUrls) or Channel names (channels)
📦 Data per messageTimestamp (HH:MM:SS), offset (s), ISO date, username, user ID, message text, emotes, emote IDs, badges, roles, color, VOD & channel metadata
🔁 Pagination & scaleParallel time-offset segmentation across the full broadcast — scales effortlessly to 1,000+ to 100,000+ chat messages per VOD with 0% duplicates
🔓 AuthenticationNone required — works without a Twitch account, API key, or OAuth token
🌐 Proxy supportStandard Datacenter proxies (enabled by default; no costly residential proxy needed)
⬇️ Export formatsBulk export to JSON, CSV, Excel (XLSX), HTML, XML, RSS, or consume via REST API & Webhooks

🔎 Why scrape Twitch VOD chat?

A Twitch stream's chat is the heartbeat of the broadcast. Every laugh, every clutch play, every clip-worthy reaction, every community meme, and every raid registers in chat seconds before anyone cuts a highlight. While live chat scrolls past in real time, Twitch preserves the full conversation as a chat replay attached to the VOD (Video on Demand / past broadcast). That replay is a goldmine of audience engagement intelligence: who spoke, at what exact second of the broadcast, what they wrote, which emotes they spammed, and how audience velocity surged and dipped across hours of streaming.

The challenge is that Twitch offers no export button, and the official Twitch API does not provide a bulk historical VOD chat replay endpoint. Scrubbing through a 6-hour or 12-hour broadcast timeline manually is impossible. This Twitch VOD chat scraper automates the entire workflow: point it at any VOD URL or channel login, and it extracts the chat replay into structured, analysis-ready rows.

High-impact ways teams use this Twitch data extractor:

  • ✂️ Automated clip & highlight discovery — plot message velocity per minute (offsetSeconds) to pinpoint the exact timestamps where chat exploded.
  • 📊 Stream engagement & audience analytics — calculate messages per minute, unique chatter counts, repeat chatter retention, and subscriber-to-non-subscriber chat ratios.
  • 😂 Emote & meme velocity tracking — rank which Twitch, subscriber, or channel emotes dominate specific segments of a broadcast.
  • 📈 Sponsorship & brand-mention attribution — measure exact chat lift, brand keyword mentions, and audience sentiment during sponsored stream segments or game reveals.
  • 🧪 Online community & academic research — study live-stream sociology, moderation patterns, VIP/subscriber dynamics, and fandom behavior across thousands of messages.
  • 🤖 Conversational AI, LLM & NLP training datasets — build authentic, timestamped multi-party chat corpora complete with emote tokens and speaker metadata.
  • 🗂️ Broadcast archiving before expiration — preserve full chat replays before Twitch automatically deletes past broadcasts (typically after 7 to 60 days).

✨ What this Twitch scraper does

  • ✅ Complete or capped VOD chat replay export — set maxMessagesPerVideo: 0 to extract the entire VOD chat from start to finish, or specify any target (200, 1,000, 10,000+) for fast representative sampling across the stream.
  • ✅ Dual targeting modes — paste exact Twitch VOD URLs (or numeric video IDs) in videoUrls, or supply channel names in channels to automatically discover and scrape their most recent past broadcasts.
  • ✅ Multi-VOD & multi-channel batching — process multiple VODs and multiple channels in a single run, controlled by maxVideosPerChannel.
  • ✅ 1,000+ deep pagination with zero duplicates — splits broadcasts into parallel time segments, advances through dense chat seconds automatically, and deduplicates every comment by its unique Twitch message ID.
  • ✅ 19 normalized fields per message — combines VOD context (videoId, videoTitle, channel, channelName, game), timing (offsetSeconds, timestamp, createdAt), chatter identity (userId, username, displayName, userColor), content (message, emotes, emoteIds), and role signals (badges, isSubscriber, isModerator, isVip).
  • ✅ Compact, spreadsheet-friendly output — empty or inapplicable optional fields are omitted per row so your CSV, Excel, and JSON exports remain clean and lightweight.
  • ✅ Zero-setup execution — runs with an empty input {} out of the box (falling back to a popular live streamer's recent VOD) so you can preview real output in seconds.

🎯 Use cases & B2B intelligence workflows

Persona / TeamWorkflow & Value Delivered
🎬 Video Editors & Clip AgenciesGroup exported rows by 30-second or 60-second windows on offsetSeconds. Sort by message count to find the top 10 loudest moments in an 8-hour VOD in under a minute—no manual watching required.
📣 Influencer Marketing & Sponsorship TeamsSearch message text for sponsor promo codes, product names, or campaign hashtags across a creator's recent broadcasts (channels + maxVideosPerChannel). Quantify live audience reaction for campaign ROI reports.
🎮 Game Publishers & Esports OrgsTrack real-time viewer sentiment during game launches, patch notes readings, or tournament finals. Compare chat velocity and top emotes across co-streamers.
📊 Creator Management & Talent AgenciesBenchmark roster creators by unique active chatters (userId), subscriber chat share (isSubscriber), and community loyalty across multiple streams.
🧠 AI / ML & Data Science TeamsIngest thousands of timestamped Twitch chat messages via API into vector databases or training pipelines for gaming-slang sentiment models and highlight-detection algorithms.

📥 Input configuration

Provide Twitch VOD URLs (videoUrls), Channel names (channels), or both. All input fields are optional—running with empty input automatically scrapes a sample broadcast so you can inspect the schema right away.

FieldTypeDefault / PrefillDescription
videoUrlsarray<string>[]Twitch VOD URLs (e.g. https://www.twitch.tv/videos/2924282282) or bare numeric video IDs (2924282282). Add as many VODs as you want.
channelsarray<string>[]Twitch channel logins or profile URLs (e.g. "kaicenat", "xqc", "pokimane"). Automatically resolves each channel's most recent past broadcasts.
maxVideosPerChannelinteger1When channels is used, how many of each channel's latest past broadcasts (archives) to scrape.
maxMessagesPerVideointegerDefault 0 (Prefill 200)Maximum chat messages to collect per VOD. Set 0 to unlock the entire chat replay (1,000+ to 100,000+ messages). Use 200–3,000 for fast sampled runs.
proxyConfigurationobject{ "useApifyProxy": true }Standard Apify Proxy configuration. Datacenter proxies are fast, cheap, and sufficient.

Example input — fast 1,000+ message export from specific VODs

{
"videoUrls": [
"https://www.twitch.tv/videos/2924282282"
],
"maxMessagesPerVideo": 1500,
"proxyConfiguration": {
"useApifyProxy": true
}
}

Example input — scrape full chat replays from recent channel broadcasts

{
"channels": [
"kaicenat",
"jynxzi"
],
"maxVideosPerChannel": 2,
"maxMessagesPerVideo": 0,
"proxyConfiguration": {
"useApifyProxy": true
}
}

📤 Output schema & field reference

Every chat message is exported as a structured object sorted chronologically by offsetSeconds. Below is a real record from a KaiCenat broadcast:

{
"videoId": "2924282282",
"videoTitle": "🎮 PLAYING EVERY GTA UNTIL I BEAT IT: Grand Theft Auto 4🎮",
"channel": "kaicenat",
"channelName": "KaiCenat",
"game": "Grand Theft Auto IV",
"offsetSeconds": 359,
"timestamp": "05:59",
"createdAt": "2026-11-09T01:45:52.735Z",
"userId": "1268652913",
"username": "james1263769",
"displayName": "james1263769",
"message": "WW STREAM",
"userColor": "#1E90FF",
"badges": "subscriber, premium",
"isSubscriber": true
}

When a message includes Twitch or channel emotes, emotes and emoteIds are automatically populated:

{
"videoId": "2924282282",
"videoTitle": "🎮 PLAYING EVERY GTA UNTIL I BEAT IT: Grand Theft Auto 4🎮",
"channel": "kaicenat",
"channelName": "KaiCenat",
"game": "Grand Theft Auto IV",
"offsetSeconds": 1559,
"timestamp": "25:59",
"createdAt": "2026-11-09T02:05:52.110Z",
"userId": "884120941",
"username": "jfrumthev",
"displayName": "jfrumthev",
"message": "el4ourYbjig el4ourYbjig",
"emotes": "el4ourYbjig, el4ourYbjig",
"emoteIds": "emotesv2_428f000bf568462cb8a3b42481be1a4d, emotesv2_428f000bf568462cb8a3b42481be1a4d",
"userColor": "#FF0000",
"badges": "subscriber",
"isSubscriber": true
}

🎬 VOD & Stream Context Fields

FieldTypeCoverageDescription
videoIdstring🟢 High (100%)Unique numeric Twitch VOD identifier.
videoTitlestring🟢 High (100%)Full title of the past broadcast or highlight.
channelstring🟢 High (100%)Broadcaster's lowercase Twitch login handle.
channelNamestring🟢 High (100%)Broadcaster's formatted display name.
gamestring🟢 High (~98%)Category or game assigned to the VOD (present whenever set on the broadcast).

⏱️ Timing & Chronology Fields

FieldTypeCoverageDescription
offsetSecondsinteger🟢 High (100%)Seconds elapsed since the start of the broadcast—ideal for time-series charting and clip detection.
timestampstring🟢 High (100%)Human-readable stream clock (MM:SS under 1 hour, HH:MM:SS over 1 hour).
createdAtstring🟢 High (100%)Exact UTC timestamp (ISO 8601) when the message was originally sent in live chat.

🧑 Chatter Identity & Role Fields

FieldTypeCoverageDescription
userIdstring🟢 High (100%)Permanent numeric Twitch user ID—remains constant even if the user renames their account.
usernamestring🟢 High (100%)Chatter's lowercase login handle.
displayNamestring🟢 High (100%)Chatter's stylized display name.
userColorstring🟡 Conditional (~80%)Hex color code chosen by the chatter (e.g. #1E90FF), present when set.
badgesstring🟡 Conditional (~65%)Comma-separated badge set IDs worn by the user (subscriber, moderator, vip, premium, turbo, founder, etc.).
isSubscriberboolean🟡 ConditionalSet to true when the chatter holds a subscriber or founder badge (omitted otherwise).
isModeratorboolean🟡 ConditionalSet to true when the chatter is a channel moderator (moderator badge).
isVipboolean🟡 ConditionalSet to true when the chatter holds a channel vip badge.

💬 Message & Emote Fields

FieldTypeCoverageDescription
messagestring🟢 High (100%)Full plain-text chat message assembled from all message fragments.
emotesstring🟡 ConditionalComma-separated emote codes used in the message (present when emotes are used).
emoteIdsstring🟡 ConditionalComma-separated Twitch emote IDs (numeric or emotesv2_…). Render any emote image via https://static-cdn.jtvnw.net/emoticons/v2/{id}/default/dark/3.0.

🔁 Deep pagination & 1,000+ message scaling

Long Twitch broadcasts generate tens of thousands of chat messages across several hours. This Actor is engineered to handle both rapid previews and massive multi-hour chat exports:

  1. Interleaved parallel segment walking: The scraper divides the VOD timeline into 120-second segments and runs 10 concurrent workers across the broadcast.
  2. No artificial per-segment starvation: When you request 1,000, 5,000, or 0 (full VOD) in maxMessagesPerVideo, active workers keep pulling pages of comments across busy segments until your exact target count is reached—so quiet stream intros or "Starting Soon" screens never reduce your total output.
  3. Guaranteed deduplication: Every comment node is stored in an in-memory deduplication map keyed by Twitch's unique comment ID and sorted by offsetSeconds before being pushed to the dataset in 500-row batches.

⚙️ How to use & integrate

  1. Open the Actor in Apify Console.
  2. Add your targets: Paste one or more VOD links into Twitch VOD URLs, or type streamer handles into Channels.
  3. Choose your volume: Leave Max messages per VOD at 200 for a quick test, set 1000–5000 for a deep sample, or set 0 to export the entire broadcast chat replay.
  4. Click Start and watch messages populate the dataset tab in real time.
  5. Download or connect: Export to JSON, CSV, or Excel, or pull the dataset via the Apify REST API, Python/Node.js SDKs, Make, Zapier, n8n, Google Sheets, or webhooks.

API example (Node.js)

import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('haketa/twitch-vod-chat-scraper').call({
channels: ['kaicenat'],
maxVideosPerChannel: 1,
maxMessagesPerVideo: 1000,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(`Exported ${items.length} timestamped Twitch chat messages.`);

API example (Python)

from apify_client import ApifyClient
client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("haketa/twitch-vod-chat-scraper").call(run_input={
"videoUrls": ["https://www.twitch.tv/videos/2924282282"],
"maxMessagesPerVideo": 2000,
})
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
print(f"[{row.get('timestamp')}] {row.get('username')}: {row.get('message')}")

💡 Pro tips for Twitch chat analysis

  • 🔥 Find clip-worthy moments in Excel or Pandas: Group rows by Math.floor(offsetSeconds / 60) to get message counts per minute. Any minute with 3× the stream's median volume is a guaranteed highlight.
  • 😂 Filter for specific reactions: Search message for terms like LUL, OMEGALUL, CLIP IT, WW, or ? spikes to classify the type of moment (funny, clutch, hype, or confusion).
  • 👑 Analyze subscriber vs. general chat: Filter on isSubscriber: true to see how paying community members react compared to raid or first-time viewers.
  • ⏳ Archive VODs before Twitch deletes them: Twitch removes past broadcasts after 7 days (Affiliates) to 60 days (Partners/Turbo). Schedule this Actor weekly on your favorite channels to build a permanent historical archive.

⚠️ Honest limitations

  • Public, non-expired VODs only: Once a past broadcast expires and is deleted from Twitch, its chat replay is permanently removed and cannot be recovered.
  • Subscriber-only VOD archives: If a broadcaster restricts VOD viewing strictly to paying channel subscribers, public metadata and chat replay may be unavailable without authentication.
  • Live streams in progress: For best results, run the scraper on completed past broadcasts (videos/<id>), or on a live VOD after it has accumulated archive segments.

❓ Frequently asked questions (FAQ)

Do I need a Twitch account, Client ID, or OAuth token? No. The Actor works out of the box with zero credentials or setup required.

How many chat messages can I scrape from a single Twitch VOD? There is no artificial ceiling. Set maxMessagesPerVideo: 0 to collect the entire VOD chat replay from start to finish—whether that is 500 messages on a short stream or 100,000+ messages on a marathon broadcast.

Why did my test run finish with 200 messages? The Console prefill sets maxMessagesPerVideo: 200 so your first test run finishes in just a few seconds and uses minimal platform compute. Change maxMessagesPerVideo to 1000, 5000, or 0 (unlimited) for full-scale extraction.

Do I need residential proxies? No. Standard Apify Datacenter proxies (enabled by default) work reliably and keep run costs very low.

Can I pass just a streamer's username instead of looking up VOD URLs? Yes! Enter one or more usernames in the channels field (for example ["kaicenat", "pokimane"]) and set maxVideosPerChannel to the number of recent broadcasts you want to scrape per streamer.

Are messages guaranteed to be in chronological order with no duplicates? Yes. All collected messages are deduplicated by their unique Twitch comment ID and sorted by offsetSeconds in ascending stream order before export.

Can I render the actual emote images from the exported data? Yes. Every emote in a message has its ID listed in emoteIds. Insert that ID into https://static-cdn.jtvnw.net/emoticons/v2/{id}/default/dark/3.0 to get the high-resolution PNG/GIF directly from Twitch's CDN.


🔌 Integrations & automation workflows

Because every run stores its output in a standard Apify Dataset, you can connect this Twitch VOD chat scraper directly into your existing analytics, content, or AI stack without writing glue code:

  • 📊 Google Sheets & Airtable: Automatically append top chat moments or per-stream summary rows after every broadcast using the native Apify Google Sheets integration.
  • ⚡ Make, Zapier & n8n: Trigger an automated workflow (Webhook → Slack / Discord / Notion) whenever a watched streamer finishes a broadcast—sending your editing team an immediate list of high-velocity timestamp spikes.
  • 🗄️ BigQuery, Snowflake & PostgreSQL: Stream raw message rows into your data warehouse to track creator growth, unique active chatters, and sponsor mention share over months of broadcasts.
  • 🤖 AI Agents & MCP (Model Context Protocol): Expose this Actor as a tool for LLM agents so they can fetch a Twitch VOD's chat replay and summarize community reaction, top questions asked in chat, and viewer sentiment on demand.

📊 Quick Pandas & SQL recipes for highlight detection

Once you export your Twitch chat dataset as CSV or JSON, you can surface the exact clip-worthy moments of a multi-hour stream in just a few lines of Python:

import pandas as pd
# Load exported CSV or JSON from Apify Dataset
df = pd.read_json("twitch_chat_dataset.json")
# 1. Find the top 10 loudest 1-minute windows (clip-worthy hype spikes)
df["minute"] = (df["offsetSeconds"] // 60).astype(int)
hype_minutes = (
df.groupby("minute")
.agg(
messages=("message", "count"),
unique_chatters=("userId", "nunique"),
sample_timestamp=("timestamp", "first"),
)
.sort_values("messages", ascending=False)
.head(10)
)
print("Top 10 Hype Minutes to Clip:")
print(hype_minutes)
# 2. Rank the most active chatters in the broadcast
top_chatters = (
df.groupby(["userId", "username"])
.size()
.reset_index(name="message_count")
.sort_values("message_count", ascending=False)
.head(15)
)
print(top_chatters)

Equivalent SQL query (BigQuery / Postgres / DuckDB)

SELECT
FLOOR(offsetSeconds / 60) AS stream_minute,
MIN(timestamp) AS start_timestamp,
COUNT(*) AS total_messages,
COUNT(DISTINCT userId) AS unique_chatters,
COUNTIF(isSubscriber) AS subscriber_messages
FROM twitch_vod_chat
GROUP BY stream_minute
ORDER BY total_messages DESC
LIMIT 15;

  • Steam Game Scraper — extract Steam store games, pricing, review scores, genres, developers, and release dates.

📋 Changelog

  • v1.2 (2026-11) — Unlocked 1,000+ to unlimited deep VOD chat pagination: switched maxMessagesPerVideo schema default to 0 (full VOD) with a fast 200-message prefill, added interleaved segment walking without per-segment starvation, added maxItems / maxMessages input aliases, and batched dataset writes for large marathon streams.
  • v1.1 (2026-10) — Added channel-to-VOD resolution (channels + maxVideosPerChannel), derived role flags (isSubscriber, isModerator, isVip), and emote ID extraction.
  • v1.0 (2026-09) — Initial release of Twitch VOD Chat Scraper with parallel offset walking and deduplication.

▶️ Ready to export Twitch chat? Paste a Twitch VOD URL or channel name, click Start, and download the complete timestamped chat replay in JSON, CSV, or Excel!