Video transcript API: one call for TikTok, YouTube, Instagram

Three platforms, three different walls. YouTube hides its captions behind OAuth, Instagram exposes none at all, and TikTok has no official API.

Build it yourself and you maintain three scrapers, three auth schemes, and three output shapes.

That's three integrations to write and keep alive, for one feature: a transcript.

One request shape returns the spoken text from any of the three, under one envelope. The rest of this page walks through it.

One shape, every platform

The call is the same on all three. You pass a URL, and the transcript comes back under output.data. Here's the call I actually run, using YouTube as the example:

curl -X POST https://api.getanyapi.com/v1/run/youtube.video_transcript \
  -H "Authorization: Bearer $ANYAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ"}'

import os, requests

res = requests.post(
    "https://api.getanyapi.com/v1/run/youtube.video_transcript",
    headers={"Authorization": f"Bearer {os.environ['ANYAPI_KEY']}"},
    json={"url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ"},
)
print(res.json()["output"]["data"]["transcript"])

const res = await fetch("https://api.getanyapi.com/v1/run/youtube.video_transcript", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.ANYAPI_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({ url: "https://www.youtube.com/watch?v=dQw4w9WgXcQ" }),
});
const { output } = await res.json();
console.log(output.data.transcript);

You change only the SKU per platform: tiktok.video_transcript for TikTok, instagram.media_transcript for Instagram. Everything else stays the same.

One key, one auth header, one envelope across all three. No Google Cloud project for YouTube, no Graph API for Instagram, no headless browser for TikTok.

What each platform gives you

The calls match; the platforms don't. How much each one exposes decides what you get back, and it's worth knowing before you build:

TikTok and YouTube publish a caption track, so you get timestamps for free. Instagram publishes none, so the audio is transcribed and you get clean text without offsets.

Price barely moves between them: $1.20 per 1,000 transcripts on TikTok and Instagram, $1.30 on YouTube.

The deep dive per platform

Each platform has its own gotcha and its own deep dive. Start with the one you need:

• TikTok. No official API, so a scraper or a hosted call is the only way. How to get a TikTok transcript with one API call.
• YouTube. The official captions.download needs OAuth and 403s on videos you don't own, and the popular OSS library breaks on cloud IPs. YouTube transcript API, explained.
• Instagram. No transcript resource and no caption track at all, so the audio has to be transcribed. Instagram transcript API, explained.

Normalize once, branch where it matters

Because the envelope is shared, a small wrapper handles every platform and you branch only on the output shape. Here's the whole seam:

async function transcript(sku, url) {
  const res = await fetch(`https://api.getanyapi.com/v1/run/${sku}`, {
    method: "POST",
    headers: {
      Authorization: `Bearer ${process.env.ANYAPI_KEY}`,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({ url }),
  });
  const { output } = await res.json();
  if (!output.found) return null;            // no transcript available
  // YouTube/TikTok: output.data.transcript ; Instagram: output.data.transcripts[].text
  return output.data.transcript ?? output.data.transcripts?.map((t) => t.text).join("\n\n");
}

Write the fetch once, and adding a platform is just a new SKU string. That's the point of a single gateway: the surface you maintain doesn't grow with the number of sources.

Limits worth knowing before you build

One envelope hides the plumbing, not the platforms. Three limits still reach your code:

• No timestamps on Instagram. There's no caption track to read, so the audio is transcribed and you get clean text with no offsets. Anything that needs to seek to a moment in the video, like clipping or a subtitle file, works on TikTok and YouTube only.
• No language selector anywhere. None of the three endpoints takes a language input, so you get whatever track the platform serves for that video. YouTube always tags what it returned, TikTok tags it when it knows, Instagram doesn't tag it at all.
• A video you can't watch has no transcript. Private, deleted, and region-locked URLs come back as output.found false with output.data null rather than an error, which is why the wrapper checks found before it touches the data.

Try them without writing code

Each platform has a free, no-login tool. Paste a link, copy the text, see the output shape before you wire up the API:

• TikTok transcript tool
• YouTube transcript tool
• Instagram transcript tool

Frequently asked questions

How much does a transcript cost?

$0.0012 per call on TikTok and Instagram and $0.0013 on YouTube, billed in dollars, with no subscription. That's $1.20 to $1.30 per 1,000 transcripts.

Can I get a transcript for free?

Yes, for a one-off. Each platform has a free in-browser tool (linked above) with no key or login. For bulk or programmatic use, the endpoint is one call at $0.0012 to $0.0013.

Can I transcribe many videos at once?

Yes. Loop the endpoint over a list of URLs; each call is independent and priced the same. The per-platform guides show the shell pattern.

What can I build with it?

Cross-platform search over a creator's whole footprint, an LLM summarizer or repurposing tool, subtitles, translation, or a RAG index for an assistant, all from one integration.

Related guides