video_id stringlengths 11 11 | title stringlengths 5 110 | channel stringclasses 2
values | upload_date timestamp[s]date 2011-12-02 00:00:00 2026-07-24 00:00:00 | url stringlengths 43 43 | duration_sec float64 18.6 15.4k | view_count null | speakers unknown | segments listlengths 0 3.73k |
|---|---|---|---|---|---|---|---|---|
--7xUR5amUM | Vita Sidorkina: Russian Model Gets Sucked Into Western Malnutrition | goatis | 2022-05-26T00:00:00 | https://www.youtube.com/watch?v=--7xUR5amUM | 944.4 | null | {
"SPEAKER_00": {
"talk_sec": 479,
"goatis_score": 0.8431000000000001
},
"SPEAKER_01": {
"talk_sec": 18.3,
"goatis_score": 0.1008
},
"SPEAKER_02": {
"talk_sec": 11.9,
"goatis_score": 0.0814
},
"SPEAKER_03": {
"talk_sec": 263.8,
"goatis_score": 0.1668
},
"SPEAKER_04": {
... | [
{
"start": 0.19,
"end": 4.56,
"text": "Hey guys, I'm going to review what a Victoria's Secret model eats in a day.",
"speaker": "SPEAKER_00",
"words": [
[
"Hey",
0.19,
0.35000000000000003,
"SPEAKER_00",
0.7000000000000001
],
[
"gu... |
--WgQNhM-oc | Why I'm No Longer Vegan - My Ex-Vegan Story | sv3rige | 2018-01-18T00:00:00 | https://www.youtube.com/watch?v=--WgQNhM-oc | 410.5 | null | {
"SPEAKER_00": {
"talk_sec": 289.1,
"goatis_score": 0.7542
}
} | [
{
"start": 0.31,
"end": 6.26,
"text": "Alright everyone, it's time to tell my ex-vegan story and why I make these anti-vegan videos.",
"speaker": "SPEAKER_00",
"words": [
[
"Alright",
0.31,
0.77,
"SPEAKER_00",
0.6900000000000001
],
[
... |
-2h5odfasHg | Pippin Jardine: Sugar & Anti-Nutrients - The Perfect Vegan Teen Diet | goatis | 2019-05-16T00:00:00 | https://www.youtube.com/watch?v=-2h5odfasHg | 387.4 | null | {"SPEAKER_00":{"talk_sec":174.0,"goatis_score":0.7708},"SPEAKER_01":{"talk_sec":113.6,"goatis_score"(...TRUNCATED) | [{"start":0.29,"end":1.47,"text":"What's up everyone?","speaker":"SPEAKER_00","words":[["What's",0.2(...TRUNCATED) |
-5E-HfgFtHI | 99% of People Are Retarded | sv3rige | 2018-10-24T00:00:00 | https://www.youtube.com/watch?v=-5E-HfgFtHI | 441.1 | null | {"SPEAKER_00":{"talk_sec":221.1,"goatis_score":0.7468},"SPEAKER_01":{"talk_sec":54.3,"goatis_score":(...TRUNCATED) | [{"start":0.15,"end":1.41,"text":"Hi, what's up?","speaker":"SPEAKER_00","words":[["Hi,",0.15,0.5700(...TRUNCATED) |
-7rAHLUGqa4 | Raw Meat Eater Reviews: Freaky Eaters | Raw Meat Addict | sv3rige | 2019-06-15T00:00:00 | https://www.youtube.com/watch?v=-7rAHLUGqa4 | 818.2 | null | {"SPEAKER_00":{"talk_sec":223.0,"goatis_score":0.6501},"SPEAKER_01":{"talk_sec":117.2,"goatis_score"(...TRUNCATED) | [{"start":0.19,"end":0.91,"text":"What's up?","speaker":"SPEAKER_00","words":[["What's",0.19,0.57000(...TRUNCATED) |
-8dJMKyYwqY | Why are Vegans Obsessed With Food, Blood and Death? | sv3rige | 2017-06-17T00:00:00 | https://www.youtube.com/watch?v=-8dJMKyYwqY | 781 | null | {"SPEAKER_00":{"talk_sec":39.5,"goatis_score":0.1262},"SPEAKER_01":{"talk_sec":36.4,"goatis_score":0(...TRUNCATED) | [{"start":0.25,"end":0.69,"text":"What's up?","speaker":"SPEAKER_05","words":[["What's",0.25,0.49,"S(...TRUNCATED) |
-8rmsFy9v0o | Ronnie Coleman & the Police: The Biggest Loser Collaboration in History @Ronnie Coleman | goatis | 2024-06-03T00:00:00 | https://www.youtube.com/watch?v=-8rmsFy9v0o | 626.5 | null | {"SPEAKER_00":{"talk_sec":382.1,"goatis_score":0.7571},"SPEAKER_01":{"talk_sec":55.1,"goatis_score":(...TRUNCATED) | [{"start":0.17,"end":1.47,"text":"Hey guys, what's up?","speaker":"SPEAKER_00","words":[["Hey",0.17,(...TRUNCATED) |
-9tFCaAVxBI | Should I Drink Pasteurized Milk If I Can't Get Raw Milk? | sv3rige | 2017-09-17T00:00:00 | https://www.youtube.com/watch?v=-9tFCaAVxBI | 113.8 | null | {
"SPEAKER_00": {
"talk_sec": 97.3,
"goatis_score": 0.84
}
} | [{"start":0.27,"end":1.03,"text":"Alright, what up?","speaker":"SPEAKER_00","words":[["Alright,",0.2(...TRUNCATED) |
-EK41BLEY_A | Be Careful Who You Trust on TikTok | goatis | 2026-03-04T00:00:00 | https://www.youtube.com/watch?v=-EK41BLEY_A | 1,292.1 | null | {"SPEAKER_00":{"talk_sec":724.9,"goatis_score":0.9148000000000001},"SPEAKER_01":{"talk_sec":41.2,"go(...TRUNCATED) | [{"start":0.21,"end":5.43,"text":"Hi, I'm going to be reviewing random TikTok videos and YouTube sho(...TRUNCATED) |
-Gy0qEGM95c | Everything That the Government Gives You for Free Is Bad for You | goatis | 2025-01-23T00:00:00 | https://www.youtube.com/watch?v=-Gy0qEGM95c | 96.9 | null | {
"SPEAKER_00": {
"talk_sec": 81.3,
"goatis_score": 0.8049000000000001
}
} | [{"start":0.27,"end":3.95,"text":"Everything that the government gives you for free is bad for you."(...TRUNCATED) |
End of preview. Expand in Data Studio
Goatis / Sv3rige Video Transcripts
Full transcripts of 1,383 videos (~487 hours, ~3.9M words) from the YouTube channels of Goatis (Sv3rige) — the sv3rige channel (2011–2025) and the current Goatis channel (2019–2026). This is the dataset behind goatis.net, a searchable archive in the style of aajonus.net.
What makes it more than raw ASR
Every video was processed with speaker identification, not just
transcription. He mostly reacts to other people's videos, so a naive transcript
attributes other people's words to him. Here, every diarized speaker cluster
carries a goatis_score — cosine similarity of its ECAPA voice embedding
against reference centroids built from five known-solo videos spanning
2012–2026. Words from reacted-to videos are separable from his own speech.
Schema (videos.jsonl.gz, one row per video)
| field | description |
|---|---|
video_id |
YouTube id — https://youtube.com/watch?v=<id> |
title, channel, upload_date, url, view_count |
video metadata |
duration_sec |
audio duration |
speakers |
per diarized speaker: talk_sec, goatis_score (cosine, ~0.55+ = him) |
segments[] |
start, end, text, speaker, words[] |
segments[].words[] |
[word, start, end, speaker, align_score] — word-level timestamps |
Pipeline
WhisperX (Whisper large-v3, float16, English) → wav2vec2 forced alignment
(word timestamps) → pyannote 3.1 diarization → speechbrain ECAPA embeddings vs
reference centroids. Processed on 2× RTX 5090 in ~18 GPU-hours. Total compute
cost of the corpus: about $3.
Usage
from datasets import load_dataset
ds = load_dataset("json", data_files="hf://datasets/exoarbuus/goatis-transcripts/videos.jsonl.gz")
# his own words only, per video
for row in ds["train"]:
his = {s for s, v in row["speakers"].items() if v["goatis_score"] >= 0.55}
text = " ".join(seg["text"] for seg in row["segments"] if seg["speaker"] in his)
Caveats
- ASR output: proper nouns are the weak point (e.g. "Aajonus" is often mangled in speech); numbers and common speech are reliable.
goatis_scoreis per cluster, not per word; overlapping speech and brief interjections can be misattributed. Scores near the 0.55 threshold are genuinely ambiguous — treat them as such.- ~46 videos are music/compilations where no speaker matches him (correctly).
- Sarcasm, quoted speech, and read-aloud comments are his voice and are attributed to him: vocal attribution ≠ endorsement of the words.
Provenance & license
Transcripts of publicly posted YouTube videos, provided for archival and research use. Underlying speech remains the speaker's. No warranty of transcription accuracy; verify against the linked timestamps before quoting.
- Downloads last month
- 19