# Changelog of Twitch VOD Chat Scraper — Full Chat Replay Export (`haketa/twitch-vod-chat-scraper`) Actor

- **URL**: https://apify.com/haketa/twitch-vod-chat-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/haketa/twitch-vod-chat-scraper.md

## Changelog

### 1.0.0 — 2026-11-01

- **Verification run:** collected **540 unique chat messages across 3 live multi-hour VODs** (KaiCenat, xQc, Jynxzi) in **~72 seconds** with **0 duplicates** and **100% coverage of all core fields** (`videoId`, `videoTitle`, `channel`, `channelName`, `game`, `offsetSeconds`, `timestamp`, `createdAt`, `userId`, `username`, `displayName`, `message`).
- **Confirmed deep pagination:** sampled messages spanned stream offsets from `7s` out to `21,478s` (~6 hours into one broadcast), proving the parallel time-segment walker drains the whole VOD, not just the opening minutes.
- **README refreshed to the gold standard:** replaced the output examples with real rows from the verification run, added a dedicated emotes example, and corrected the `emoteIds` documentation to cover both legacy-numeric (`64138`) and modern (`emotesv2_…`) IDs (both verified against the Twitch CDN image pattern). Documented honest, data-dependent field coverage (`userColor` ≈ 79%, `badges` ≈ 65%, `emotes` ≈ 4%).
- No code changes required — the scraper is healthy: GQL transport, client-ID rotation, datacenter-proxy rotation, retries and the comment-ID dedup sink all behaving correctly.
