For AI
View as MarkdownOpen in ChatGPTOpen in Claude

YouTube video transcript API

Get the transcript of a YouTube video by ID or link: the whole text and timed segments. One object a call; $0.0020 per call. Errors, timeouts and objects that do not exist are not charged.

GEThttps://api.dataxapi.com/v1/youtube/videos/transcript/get
Also known asyoutube captionsvideo subtitlesYouTube video captionsIn Chinese: YouTube 视频字幕 API
id is in the link: youtube.com/watch?v=<id>. Or pass the whole link as url.No code: try the YouTube transcript extractor →
Price$0.0020/ call · Errors and timeouts are not charged
Returns1 objectone transcript object
When an upstream failsRetried for youon the same route, or a backup route; never charged
StatusLive statusChecked daily with real calls · 4 days on record

Try it

Change a value and the request on the right follows.
Run · $0.0020

Sign in to run it; it runs on your own balance. New accounts can try this endpoint free, up to 50 calls.

Used to call POST /v1/youtube/videos/captions? It is this address now. What changed for old endpoints

The first 2 rows of a response recorded 2026-10-12 03:34 UTC.
startend
0.44.24
2.086.56
curl -G "https://api.dataxapi.com/v1/youtube/videos/transcript/get" \
  --data-urlencode "id=dQw4w9WgXcQ" \
  -H "Authorization: Bearer $DATAXAPI_API_KEY"
{
  "code": "ok",
  "request_id": "45a3a507-98c9-484f-a823-690b541e8c9d",
  "data": {
    "item": {
      "post_id": "pW1QdpFNg3a",
      "lang": "en",
      "text": "Hey everyone, today I am going through what I packed for weekend hiking. Short version: light, fits in one bag, hard to break. Let's go piece by piece.",
      "segments": [
        {
          "start": 0.4,
          "end": 4.24,
          "text": "Hey everyone, today I am going through what I packed for weekend hiking."
        },
        {
          "start": 2.08,
          "end": 6.56,
          "text": "Short version: light, fits in one bag, hard to break."
        },
        // … 61 more
      ]
    }
  },
  "warnings": [],
  "meta": {
    "capability": "youtube.videos.transcript.get",
    "result_status": "complete",
    "fetched_at": "2026-10-12T03:34:28.000Z",
    "cached": false,
    "replayed": false,
    "unavailable": {},
    "billing": {
      "amount": "0.002",
      "currency": "USD",
      "balance_after": "9.998"
    }
  }
}

From a response recorded 2026-10-12 03:34 UTC. Names, text and IDs are examples, and so are the links: none is the recorded one (the page shows no one's personal data); the figures are as recorded. Long lists are cut. The request ID, cursor and balance are made up.

Size as sent: 6.9 KB in the standard view (931 B compact, 6.9 KB full).

This example as a JSON file

curl -G "https://api.dataxapi.com/v1/youtube/videos/transcript/get" \
  --data-urlencode "id=dQw4w9WgXcQ" \
  -H "Authorization: Bearer $DATAXAPI_API_KEY"
from dataxapi import DataXAPI

dx = DataXAPI()  # reads DATAXAPI_API_KEY
result = dx.youtube.videos.transcript.get(id="dQw4w9WgXcQ")
print(result["data"]["item"])
import { DataXAPI } from "@dataxapi/sdk";

const dx = new DataXAPI(); // reads DATAXAPI_API_KEY
const result = await dx.youtube.videos.transcript.get({ id: "dQw4w9WgXcQ" });
console.log(result.data.item);
dataxapi call youtube.videos.transcript.get id=dQw4w9WgXcQ
{
  "mcpServers": {
    "dataxapi": {
      "type": "http",
      "url": "https://api.dataxapi.com/mcp",
      "headers": {
        "Authorization": "Bearer ${DATAXAPI_API_KEY}"
      }
    }
  }
}

Add this to your MCP client, then ask your agent to call youtube.videos.transcript.get.

Parameters

idtextgive exactly one of id, url

Video ID. Example: dQw4w9WgXcQ.

urltextgive exactly one of id, url

Link to the video: a watch, shorts, embed, live or youtu.be link. Example: https://www.youtube.com/watch?v=dQw4w9WgXcQ. Full link: https://www.youtube.com/watch?v=dQw4w9WgXcQ. Short link: https://youtu.be/dQw4w9WgXcQ.

viewtext

How much of each item to return. All three cost the same. Values: compact, standard (default), full.

fieldslist of text

Return only these fields. Each by its path, such as author.name.

Known limits

  • One object per call: there is nothing to page.

Returned fields

data.item is one transcript object. Stable: fields are only ever added.

post_idtextcompact

ID of the post it belongs to

langtextcompact

Language code

texttextcompact

The whole text

Long text: the compact view keeps the first 500 characters and says so in warnings.

segmentslist of objectstandard

Timed segments

segments[].startnumberstandard

Start, in seconds

segments[].endnumberstandard

End, in seconds

Show all 10 fields →
segments[].texttextstandard

Text of the segment

approximatelist of objectcompact

Which numbers of this item are approximate or lower bounds; absent when all are exact

approximate[].fieldtextcompact

Path of the field, such as metrics.views

approximate[].kindtextcompact

Approximate, or a lower boundapproximatelower_bound

Typical size of one item: compact 552 B, standard 6.6 KB, full 6.6 KB (measured on the response recorded 2026-10-12 03:34 UTC).

When you are charged

A result with data$0.0020, once per call
The object does not exist (404)not charged
Invalid parameters, a failure of ours or of the data source, a timeoutnot charged
CaseCharged
The most one call can cost$0.0020 (send X-Max-Price to refuse anything above your own limit)

Prices as read 2026-10-21 09:10 UTC, under price book 2, in effect since 2026-10-17 11:00 UTC; the Playground reads the price in force before a run.

If the object does not exist

The answer is 404 not_found and nothing is charged. reason says which: not_found (it does not exist) or private_or_deleted.

Errors

What this endpoint can answer with; codes all endpoints share are on the full list.

HTTPCodeMeaningCharged
400invalid_paramsA parameter is invalidRetry: After a fixNo
402insufficient_balanceThe balance is too lowRetry: After a fixNo
404not_foundThe object does not exist, is private or was deletedRetry: NoNo
502upstream_failedThe data source failedRetry: YesNo
504timeoutThe data did not arrive in time; retry with the same idempotency keyRetry: YesNo

All error codes →

Questions

How much does the YouTube video transcript API cost?

$0.0020 per call. The most one call can cost is $0.0020.

When am I not charged?

Errors, timeouts and objects that do not exist are not charged.

What happens if the object does not exist?

The answer is 404 not_found and nothing is charged. reason says which: not_found (it does not exist) or private_or_deleted.

The same thing on other platforms

Meta ad video transcript$0.0020Instagram video transcript$0.0020TikTok video transcript$0.0020LinkedIn video transcript$0.0020Reddit video transcript$0.0020X video transcript$0.0020

All 10 YouTube endpoints

YouTube endpoints
YouTube channel profile$0.0020YouTube channel search$0.0020YouTube channel videos$0.0020YouTube comment replies$0.0020YouTube video comments$0.0020YouTube video detail$0.0020YouTube video search$0.0020YouTube video transcript$0.0020YouTube channel posts$0.0020YouTube post detail$0.0020

Use it in your AI

Copy this prompt

Your API key is built in, so there is nothing to configure.

Read https://dataxapi.com/agents.md and set up DataXAPI in this agent. My API key is dx_…; use it for every DataXAPI call.
Sign in to copy it with your key

No account yet? Sign-up takes an email and no card, and a failed call is never charged.

then paste

Paste it into your agent

It reads the page the prompt points to and connects itself.

Your agent
Read https://dataxapi.com/agents.md and set up DataXAPI in this agent. My API key is dx_••••; use it for every DataXAPI call.
DataXAPI connected · 5 tools ready
now try it

Now ask something

Use one of these, or your own words.

Read https://dataxapi.com/apis/youtube/videos/transcript/get.md and help me use the DataXAPI endpoint youtube.videos.transcript.get (YouTube video transcript).
Prefer to set it up by hand?
dataxapi setupgives the agents on your machine (Claude Code, Codex, Cursor) the DataXAPI MCP server and the base skill. The key is read from DATAXAPI_API_KEY; none is written into a file.MCP server: https://api.dataxapi.com/mcp. The entry for your client is under MCP in the request examples of this page; then ask your agent to call youtube.videos.transcript.get.Other clients →
Ask with this pageSends the Markdown version
Open in ChatGPTOpen in ClaudeOther clients →
Brief for an AIEverything needed to call it, as text
Show the text
Use the DataXAPI endpoint youtube.videos.transcript.get (YouTube video transcript). Get the transcript of a YouTube video by ID or link: the whole text and timed segments.
Call GET https://api.dataxapi.com/v1/youtube/videos/transcript/get with the header "Authorization: Bearer $DATAXAPI_API_KEY". Read the key from the environment variable DATAXAPI_API_KEY; never write it into code.
Parameters:
- id: Video ID. Example: dQw4w9WgXcQ.
- url: Link to the video: a watch, shorts, embed, live or youtu.be link. Example: https://www.youtube.com/watch?v=dQw4w9WgXcQ. Full link: https://www.youtube.com/watch?v=dQw4w9WgXcQ. Short link: https://youtu.be/dQw4w9WgXcQ.
- view: How much of each item to return. All three cost the same. Values: compact, standard (default), full.
- fields: Return only these fields. Each by its path, such as author.name.
Price: $0.0020 per call. Errors, timeouts and objects that do not exist are not charged.
The answer is JSON: data.item holds transcript; meta.billing.amount is what this call was charged.
Example:
curl -G "https://api.dataxapi.com/v1/youtube/videos/transcript/get" \
  --data-urlencode "id=dQw4w9WgXcQ" \
  -H "Authorization: Bearer $DATAXAPI_API_KEY"
Full reference in Markdown: https://dataxapi.com/apis/youtube/videos/transcript/get.md

Names and identifiers

In Chinese
YouTube 视频字幕 API
Endpoint ID
youtube.videos.transcript.get
Endpoint
GET /v1/youtube/videos/transcript/get
Last checked 2026-10-21 09:10 UTC.View the Markdown versionopenapi.jsonllms.txt

Start with one key.

207 endpoints on 21 platforms, from $0.0003 a call. Errors and timeouts are not charged.

Start free