TikTok has no "Show transcript" button. Unlike YouTube, there's no panel to open and nothing to copy — the auto-captions are built for watching, not exporting — and a large share of videos have no captions at all. So if you need the spoken text of a TikTok video in your product (hook research, repurposing, search, an AI pipeline), you have to extract it. Here's the fast way, including the case most tools quietly skip: videos with no captions.

The shortest version is one request:

curl -sS \
  -H "x-api-key: $SOCIALFETCH_API_KEY" \
  -G "https://api.socialfetch.dev/v1/tiktok/videos/transcript" \
  --data-urlencode "url=https://www.tiktok.com/@mrbeast/video/7596844935442189598"

You'll need an API key and curl or the TypeScript SDK. New here? Start with the Quickstart.

Why TikTok transcripts are hard to get

YouTube hands you a transcript panel. TikTok doesn't. What you're actually up against:

• No export. Auto-captions have existed since 2021, but there's no copy-all and no download. You can watch them on screen — that's it.
• Mobile-only captions. The toggle isn't reliable on desktop web, so even manual copying is clumsy.
• Captions are optional. A huge share of videos have none — so caption-scraping returns nothing, and the only way to get text is to transcribe the audio.
• Formatted for viewing, not reuse. Inconsistent punctuation, broken sentence boundaries, and mangled proper nouns are normal.

That third point is why this is harder than it looks: getting a transcript means handling both the "captions exist" case and the "no captions, transcribe the audio" case. Most tools — free caption scrapers especially — only do the first.

You can build the DIY caption-scraper path yourself with open-source libraries, but it carries the same maintenance burden as any TikTok scraper (proxies, bot detection, breakage on every internal change) — see how to scrape TikTok data for why. This guide stays focused on getting clean text out fast.

Transcript vs captions vs subtitles

These three words get used interchangeably, but they mean different things — and which one you want changes what you do with the output:

A transcript is not the on-screen text overlays, the hashtags, or the description — it's the spoken words. The endpoint returns timestamped WebVTT, which gives you both: keep the timings for captions, or strip them for a clean plain-text transcript (shown below).

Get a transcript

Pass a TikTok video URL, get the spoken transcript back in the same data + meta envelope as the rest of the API:

The same lookup across cURL, the TypeScript SDK, Python, and more:

const videoUrl = "https://www.tiktok.com/@mrbeast/video/7596844935442189598";

const response = await fetch(
  `https://api.socialfetch.dev/v1/tiktok/videos/transcript?url=${encodeURIComponent(videoUrl)}`,
  {
    headers: {
      "x-api-key": process.env.SOCIALFETCH_API_KEY,
    },
  }
);

const body = await response.json();

console.log(response.status, body);

The only required parameter is url — the full link to a public video. Both the long share URL and vm.tiktok.com short links work. Full details are in the endpoint reference.

When a video has no captions

This is the case the manual and DIY routes can't handle: if a creator never enabled captions, there's nothing to scrape. Set useAiFallback=true and the audio is transcribed with speech-to-text instead, so you get text either way:

const videoUrl = "https://www.tiktok.com/@mrbeast/video/7596844935442189598";

const response = await fetch(
  `https://api.socialfetch.dev/v1/tiktok/videos/transcript?url=${encodeURIComponent(videoUrl)}&useAiFallback=true`,
  {
    headers: {
      "x-api-key": process.env.SOCIALFETCH_API_KEY,
    },
  }
);

const body = await response.json();

console.log(response.status, body);

The fallback adds 10 credits and only charges on a completed lookup. Leave it off when you only want existing captions and would rather skip videos without them; turn it on when you need a transcript for every video regardless of whether the creator captioned it — usually the case for research and repurposing pipelines.

Requesting a specific language

Pass an optional two-letter language code to request a specific language when one is available:

const videoUrl = "https://www.tiktok.com/@mrbeast/video/7596844935442189598";

const response = await fetch(
  `https://api.socialfetch.dev/v1/tiktok/videos/transcript?url=${encodeURIComponent(videoUrl)}&language=en`,
  {
    headers: {
      "x-api-key": process.env.SOCIALFETCH_API_KEY,
    },
  }
);

const body = await response.json();

console.log(response.status, body);

If the requested language isn't available, the lookup resolves to what TikTok has. Check data.transcript to see what came back.

Reading the response

A successful response wraps the transcript alongside the video identity and billing metadata:

{
  "data": {
    "lookupStatus": "found",
    "video": {
      "id": "7596844935442189598",
      "url": "https://www.tiktok.com/@mrbeast/video/7596844935442189598"
    },
    "transcript": {
      "format": "webvtt",
      "content": "WEBVTT\n\n00:00:00.060 --> 00:00:03.100\nThis is the world's largest LED floor.\n\n00:00:03.101 --> 00:00:09.433\nAnd now it's the world's largest green screen."
    }
  },
  "meta": {
    "requestId": "req_01example",
    "creditsCharged": 1,
    "version": "v1"
  }
}

The fields that matter:

• data.lookupStatus — found or not_found. A 200 does not guarantee found; a missing or private video returns 200 with not_found. Handle it in application logic.
• data.video — the resolved video's id and canonical url, so you can reconcile the transcript to its source.
• data.transcript.content — the raw WebVTT text (data.transcript.format is webvtt).
• meta.creditsCharged — exactly what this call cost (1, or 11 with the AI fallback), so you can budget batch jobs.
• meta.requestId — trace identifier if a transcript ever looks wrong.

From WebVTT to plain text

data.transcript.content is WebVTT, which keeps the timestamps. Use it as-is for captions; for a clean plain-text transcript — blog drafts, prompts, search indexing — strip the timing cues:

import { SocialFetchClient } from "@socialfetch/sdk";

const client = new SocialFetchClient({
  apiKey: process.env.SOCIALFETCH_API_KEY!,
});

const result = await client.tiktok.getVideoTranscript({
  url: "https://www.tiktok.com/@mrbeast/video/7596844935442189598",
});

if (!result.ok) {
  console.error(result.error.code, result.error.requestId);
  process.exit(1);
}

const vtt = result.value.data.transcript?.content ?? "";

const plainText = vtt
  .replace(/\r\n/g, "\n")
  .split("\n")
  .map((line) => line.trim())
  .filter(
    (line) =>
      line &&
      line !== "WEBVTT" &&
      !line.includes("-->") &&
      !/^\d+$/.test(line),
  )
  .join(" ");

console.log(plainText);

That gives you both outputs from a single request: the timed version for video, the flat text for everything else.

What you can build

• Hook research — transcribe a set of viral videos in a niche, pull the first line of each, and study what's actually getting watched.
• Content repurposing — turn one TikTok into a thread, newsletter section, or blog draft without rewatching it, as a "video → text → distribution" pipeline.
• AI and RAG pipelines — feed transcripts into summarization, sentiment, or retrieval. Text is searchable and composable in ways raw video never is.
• Accessibility and indexing — generate captions for your own reposts, or index spoken content so it's findable in a knowledge base.

FAQ

Does TikTok have a built-in transcript feature?

Not for export. TikTok shows auto-generated captions on screen during playback (mobile, and only if the creator enabled them), but there's no copy-all, no download, and no transcript panel like YouTube's. To get reusable text you need a tool or an API.

What if the video has no captions?

Set useAiFallback=true. The audio is transcribed with speech-to-text and you get text even when no caption track exists — the main reason an API beats DIY caption scrapers, which return nothing for uncaptioned videos. The fallback adds 10 credits on a completed lookup.

What format is the transcript returned in?

WebVTT — timestamped text. Use it directly as captions, or strip the timing cues for plain text. See From WebVTT to plain text.

Can I get transcripts in other languages?

Pass the optional two-letter language parameter when a specific language is available. If it isn't, the lookup resolves to what TikTok has — check data.transcript.

Is getting TikTok transcripts legal?

Transcribing publicly available video for analysis and repurposing is a common, widely-practiced pattern, and courts (e.g. hiQ v. LinkedIn) have generally treated public-web data as fair game in several jurisdictions. You remain responsible for TikTok's terms, copyright and privacy law, and your own contracts — a transcript is the creator's spoken words, so attribution and fair-use considerations apply when you republish. This is a technical guide, not legal advice.

How does this compare to other providers?

We keep honest, side-by-side comparisons: vs Apify, vs Bright Data, vs EnsembleData, and the full compare hub.

Next steps: Transcript endpoint reference · Scrape TikTok profiles & videos · Quickstart · Pricing