How to download the transcripts of a TikTok account's videos

What the scraper reads on an account

Without signing in, TikTok publishes about the 10 latest videos of an account, pinned ones first, in the list it serves for embedding the account on other websites. That is what the scraper reads. TikTok's full list needs signed requests, which we do not make; older videos can be transcribed from their links. TikTok's own APIs do not return speech, outside its Research API for non-profit researchers (details in our guide on the TikTok transcript API).

A text and a subtitle file per video, in Python

import csv
import pathlib
from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("quarra/tiktok-transcript-scraper").call(run_input={
    "urls": ["@nasa"],
    "maxVideosPerAccount": 10,
    "includeSrt": True,
})

folder = pathlib.Path("nasa")
folder.mkdir(exist_ok=True)
videos = list(client.dataset(run.default_dataset_id).iterate_items())
videos.sort(key=lambda v: v.get("position") or 0)  # records arrive as videos finish
with open(folder / "index.csv", "w", newline="", encoding="utf-8") as f:
    w = csv.writer(f)
    w.writerow(["videoId", "publishedAt", "durationSeconds", "viewCount", "language", "source", "wordCount", "status"])
    for v in videos:
        if v["status"] == "ok":
            (folder / f"{v['videoId']}.txt").write_text(v["transcript"], encoding="utf-8")
            (folder / f"{v['videoId']}.srt").write_text(v["srt"], encoding="utf-8")
        w.writerow([v.get("videoId"), v.get("publishedAt"), v.get("durationSeconds"), v.get("viewCount"),
                    v.get("language"), v.get("transcriptSource"), v.get("wordCount"),
                    v.get("missingReason") or v["status"]])
print(run.status, len(videos), "videos ->", folder.resolve())

Tested with apify-client 3.3 on October 10, 2026, in an 83-second call. It wrote 5 text files, 5 SRT files and an index with these rows:

The other 5 came back as free no_speech records and kept their row, with no file. On four of them, plain Whisper hears nothing or “I'm out!”; the fifth holds a single line over 25 seconds. With apify-client 2.x, write run["defaultDatasetId"] instead of run.default_dataset_id.

Two columns tell you what each transcript is worth. transcriptSource says whether it comes from TikTok's captions or from speech-to-text, and each record's isAutoGenerated whether a person wrote the captions. Captions written by the creator are usually the cleanest; TikTok's automatic ones are returned as TikTok made them, mistakes included.

Keep up with an account

Schedule the Actor on Apify with the same input and keep the videos you have not seen yet: videoId is stable, so a set of the IDs already saved is enough. A video with captions costs $0.001 whatever its length, so a daily run on a busy account stays cheap.

Limits

• About 10 latest videos per account. No hashtags, sounds, search or comments.
• The no-speech check keeps a margin: a single short line over music can come back as no_speech, free.
• TikTok sometimes answers with a bot check instead of a page, about 2 pages in 100 in our tests. We never solve those checks: the page is asked again, and only then the video gets a free record marked retryable.
• Private and age-restricted videos need a signed-in account; we do not use one, so they come back as free error records.

What it costs

$1.00 per 1,000 transcripts on Apify's free plan ($0.75 on Gold and above), plus $0.001 per started minute only when a video is transcribed by AI, and a start fee of $0.003 per run at the default memory. The 10 videos of @nasa above cost $0.008: 5 transcripts from captions and the start fee. A 90-second video without captions costs $0.003. Apify's free plan includes $5 of usage a month.