# YouTube video transcript API

> The transcript of a video with timestamps, in the original or a chosen language.

Canonical page: https://okdata.online/apis/youtube/videos/transcript/get/

- **Endpoint**: `GET /v1/youtube/videos/transcript/get`
- **ID**: `youtube.videos.transcript.get`
- **Price**: $0.004 per call
- **Returns**: one Transcript object (`data.item`)
- **Status**: preview
- **Good for**: Transcripts, Content research

## Price

One call costs $0.004 and returns one transcript. 1,000 lookups cost $4.00. A lookup that reached the data source is charged even when it is not found or the source fails; calls stopped before reaching a source are free.

## Parameters

| Name | Type | Required | Default | Description |
|---|---|---|---|---|
| `id` | string | yes |  | The video ID. (e.g. `dQw4w9WgXcQ`) |
| `lang` | string | no |  | Preferred transcript language. The result's lang is the track actually returned. (e.g. `en`) |
| `view` | string: compact / standard / full | no | `standard` | How much of each object comes back. full adds platform-specific fields under ext. |
| `fields` | string[] | no |  | Fields to keep, by path (author.name), to save tokens. In a query string: comma-separated. |

cursor, view and fields work as described in [Requests](https://okdata.online/docs/requests/).

## Returned fields

Each item is a [Transcript](https://okdata.online/docs/objects/transcript/).

| Field | Type | Description |
|---|---|---|
| `video_id` | string | ID of the video. |
| `lang` | string | Language of this transcript. |
| `segments` | Segment[] | Timed segments in order. |
| `text` | string | All segments joined as plain text. |

## Request examples

### cURL

```bash
curl -G "https://api.okdata.online/v1/youtube/videos/transcript/get" \
  -H "Authorization: Bearer $OKDATA_API_KEY" \
  --data-urlencode 'id=dQw4w9WgXcQ'
```

### Python

```python
import os
import requests

resp = requests.get(
    "https://api.okdata.online/v1/youtube/videos/transcript/get",
    headers={"Authorization": f"Bearer {os.environ['OKDATA_API_KEY']}"},
    params={"id": "dQw4w9WgXcQ"},
    timeout=60,
)
body = resp.json()
if body["code"] != "ok":
    raise RuntimeError(f"{body['code']}: {body['message']} (charged: {body['charged']})")

row = body["data"]["item"]
print(row["text"][:200])
print("cost: $" + body["meta"]["billing"]["amount"])
```

### TypeScript

```ts
const url = new URL('https://api.okdata.online/v1/youtube/videos/transcript/get');
url.searchParams.set('id', "dQw4w9WgXcQ");

const res = await fetch(url, {
  headers: {
    Authorization: `Bearer ${process.env.OKDATA_API_KEY}`,
  },
});
const body = await res.json();
if (body.code !== 'ok') throw new Error(`${body.code}: ${body.message}`);

const row = body.data.item;
console.log(row.text.slice(0, 200));
console.log('cost: $' + body.meta.billing.amount);
```

### MCP

```json
{
  "name": "call",
  "arguments": {
    "capability": "youtube.videos.transcript.get",
    "id": "dQw4w9WgXcQ"
  }
}
```

Or ask an agent that has the MCP server connected:

> Use YouTube video transcript and summarise what the video says in five bullet points.

## Response example

```json
{
  "code": "ok",
  "request_id": "0f6c2a8e-4b1d-4e0a-9a51-3c7d2b9e1f20",
  "data": {
    "item": {
      "video_id": "Zx3kP9vQm2A",
      "lang": "en",
      "segments": [
        {
          "start": 0,
          "end": 4.2,
          "text": "Here is what I pack for a two-night trip."
        }
      ],
      "text": "Here is what I pack for a two-night trip."
    }
  },
  "warnings": [],
  "meta": {
    "capability": "youtube.videos.transcript.get",
    "billing": {
      "amount": "0.004",
      "currency": "USD",
      "balance_after": "9.996"
    },
    "unavailable": {},
    "cached": false,
    "replayed": false,
    "fetched_at": "2026-11-11T08:00:00Z"
  }
}
```

## Errors

Errors share one shape: code, request_id, message and message_zh (a fixed sentence per code), charged, retry, retry_after_seconds, param and docs_url, with the specifics in problems. An error from a call that reached the data source (not found, a failure or a timeout of the source) is charged once at this endpoint's price, with charged: true and amount; errors before that, such as bad parameters, are not charged. Only settlement_unknown is unknown until looked up. The full table: https://okdata.online/docs/errors/

## Questions

### How much does the YouTube video transcript API cost?

One call costs $0.004 and returns one transcript. 1,000 lookups cost $4.00. A lookup that reached the data source is charged even when it is not found or the source fails; calls stopped before reaching a source are free.

### Do I need a YouTube account, cookies or a QR login?

No. You send only your okdata API key. The data is public data from YouTube; nothing is fetched with your own YouTube account.

### Am I charged when a call fails?

Yes, once, when the call reached the data source: an empty answer, a not_found, a failure or a timeout of the source costs one call (one item for per-item endpoints), and the error says charged: true with amount. Calls refused before reaching a source, such as bad parameters or a low balance, cost nothing.

### What if YouTube does not show a field?

The field is null, and meta.unavailable names it with the reason: platform (the platform never shows it) or capability (we cannot fetch it yet).

## More YouTube endpoints

- [YouTube video search](https://okdata.online/apis/youtube/videos/search/): $0.004 per call · about 20 videos per page
- [YouTube video detail](https://okdata.online/apis/youtube/videos/get/): $0.002 per call
- [YouTube video comments](https://okdata.online/apis/youtube/videos/comments/list/): $0.002 per call · about 20 comments per page
- [YouTube comment replies](https://okdata.online/apis/youtube/comments/replies/list/): $0.002 per call · about 10 comments per page
- [YouTube channel posts](https://okdata.online/apis/youtube/channels/posts/list/): $0.002 per call · about 10 posts per page
- [YouTube post detail](https://okdata.online/apis/youtube/posts/get/): $0.002 per call
- [YouTube channel profile](https://okdata.online/apis/youtube/channels/get/): $0.002 per call
- [YouTube channel search](https://okdata.online/apis/youtube/channels/search/): $0.004 per call · about 20 accounts per page
- [YouTube channel videos](https://okdata.online/apis/youtube/channels/videos/list/): $0.002 per call · about 30 videos per page

## At a glance

- **Short name**: YouTube video transcript
- **Category**: [Social media](https://okdata.online/categories/social/)
- **Data type**: [Transcript API](https://okdata.online/roles/transcripts-get/)
- **Status**: preview. It may still change; any change is announced beforehand.
- **Price**: $0.004 per call · bad parameters are free
- **Returns**: One Transcript, as data.item
- **If a call fails**: Charged once. When a data source fails, the next one is tried. A call that reached a source is charged once even if it fails; bad parameters are free.

## When it is charged

- Data comes back: $0.004 per call
- Bad parameters, too little balance, a paused endpoint: Free

| When | Charge |
|---|---|
| Most one call costs | $0.004 |
| An empty answer | $0.004 |
| Not found (the platform has no such item) | $0.004 |
| The data source fails or times out | $0.004 |
| Bad parameters, too little balance, a paused endpoint | Free |
| Resent with the same Idempotency-Key within 24 hours | Free when the kept result comes back |

A call that reached the data source costs $0.004 even when the answer is empty, not found or a failure of the source; a call stopped before that, such as for bad parameters, is free.

Prices from price book 84f29877acc83a88. [Every price on this platform](https://okdata.online/pricing/youtube/)

- New accounts get $0.10 of credit, no card needed: 25 calls to this endpoint.
- 1,000 lookups cost $4.00.

## Fields by view

The view column is the smallest `view` a field comes back in: `compact` returns only compact fields, `standard` adds the standard ones, `full` returns everything. The default is `standard`.

- `compact`: `video_id`, `lang`, `text`
- `standard`: `segments`

## Common errors on this endpoint

| HTTP | Code | Meaning | Charged | Retry |
|---|---|---|---|---|
| 400 | `invalid_params` | A parameter is missing, has the wrong type or value, or the method is wrong. problems lists each one; param names the first. | no | after a fix |
| 401 | `unauthorized` | The key is missing, malformed, revoked, disabled or expired (reason says which when known). | no | no |
| 402 | `insufficient_balance` | The balance (available) cannot cover the most this call could cost (required). | no | after a fix |
| 404 | `not_found` | The post, account or page does not exist or is not public. When the data source billed the lookup, or the endpoint's charge_when_not_found is above 0, this answer is charged: the body then says charged: true and amount. | once, if it reached a data source | no |
| 429 | `rate_limited` | The account sent too many requests this minute (reason: rate) or has too many calls running (reason: concurrency). | no | yes |
| 500 | `settlement_unknown` | The data came back, but the charge could not be confirmed. | unknown until looked up | with the same Idempotency-Key |
| 502 | `upstream_failed` | Every data source for this endpoint failed on this request. If a source billed the attempt, it is charged once at the endpoint's price (charged: true and amount); otherwise nothing is charged. | once, if it reached a data source | yes |
| 503 | `capability_paused` | This endpoint is paused right now; nothing ran. | no | yes |
| 504 | `timeout` | The data did not come back in time. If the request had reached a data source that bills it, it is charged once at the endpoint's price (charged: true and amount); otherwise the hold is released. | once, if it reached a data source | yes |

[All error codes](https://okdata.online/docs/errors/)

## Same data on other platforms

- [Instagram reel transcript](https://okdata.online/apis/instagram/posts/transcript/get/): $0.004 per call
- [Reddit video transcript](https://okdata.online/apis/reddit/posts/transcript/get/): $0.004 per call
- [TikTok video transcript](https://okdata.online/apis/tiktok/videos/transcript/get/): $0.002 per call

[Transcript API](https://okdata.online/roles/transcripts-get/)

## Use it in your AI

> Read https://okdata.online/apis/youtube/videos/transcript/get.md, then call this endpoint for me through okdata; my key is in the OKDATA_API_KEY environment variable. Use YouTube video transcript and summarise what the video says in five bullet points.

## Answers that use this endpoint

- [How do I download transcripts for hundreds of YouTube videos at once?](https://okdata.online/answers/bulk-youtube-transcripts/): Put the video IDs in a list and call the YouTube transcript endpoint for each one with a few workers in parallel; okdata absorbs the YouTube side, so there is no IP to get banned. Each transcript costs $0.004, 1,000 videos come to $4.00 at most, and a video without captions is charged like one with captions.
- [How do I add TikTok and YouTube data as tools in a Node.js agent?](https://okdata.online/answers/node-agent-tiktok-youtube-tools/): Wrap each endpoint in a tool with a zod schema and a fetch call: the model picks the tool, your code calls okdata and returns trimmed JSON. A TikTok search tool costs $0.002 per call for about 30 videos, and a YouTube transcript tool $0.004 per video; nothing runs on your server except the HTTP request.
- [Should I use a platform's official API or a third-party social data API?](https://okdata.online/answers/official-api-vs-third-party-social-data/): Use the official API when you need to post, need private metrics of accounts you control, or need a direct agreement with the platform. Use a third-party API for public data about accounts you do not control, or when the official one does not offer it: YouTube's API cannot download captions for videos you cannot edit, while okdata returns any public video's transcript for $0.004. X's official reads are priced per post; okdata X search is $0.002 a page of about 20.
- [How do I let Claude read and summarize a YouTube video?](https://okdata.online/answers/summarize-youtube-video-claude/): Connect the okdata MCP server to Claude; it can then fetch a video's transcript ($0.004), its details ($0.002) and comment pages ($0.002 each) by itself. A summary that also reads two pages of comments costs $0.01.
- [Is there an MCP server that gives Claude TikTok, YouTube and X data?](https://okdata.online/answers/tiktok-youtube-x-mcp-server/): Yes. Add the okdata MCP server (https://api.okdata.online/mcp) to Claude with your API key, and Claude can search TikTok ($0.002 a page), read YouTube transcripts ($0.004 a video) and search X ($0.002 a page), plus comments, profiles and trends, all with one key and no platform logins. A ten-call research session across the three costs $0.035.
- [Why does YouTube's captions.download return 403 for videos I don't own?](https://okdata.online/answers/youtube-captions-download-403/): Because captions.download requires permission to edit the video: no OAuth scope fixes a 403 on someone else's upload. For public videos you don't own, the YouTube transcript endpoint returns the captions as text with timestamps for $0.004 per video, with no OAuth and no Google Cloud project.
- [youtube-transcript-api says my IP is blocked. How do I get YouTube transcripts from a server?](https://okdata.online/answers/youtube-transcript-ip-blocked/): Stop fetching captions from YouTube with your server's own IP: call the YouTube transcript endpoint (/v1/youtube/videos/transcript/get) and okdata does the fetching. Each transcript costs $0.004, a video without captions returns not_found and costs the same, and you need no residential proxies.
- [How do I get a YouTube transcript in a specific language?](https://okdata.online/answers/youtube-transcript-language/): Pass lang (for example en or es) to the YouTube transcript endpoint; it returns that track when the video has one, and the response carries a lang field you can check before storing the text. Each transcript costs $0.004, so check lang before asking again, and translate the segments yourself when the video only has another language.
- [How do I get a YouTube transcript with timestamps as JSON or SRT?](https://okdata.online/answers/youtube-transcript-timestamps-srt/): The YouTube transcript endpoint returns every caption line as a segment with start and end in seconds, plus the full text, for $0.004 per video. Keep the JSON as it is, or write the segments out as SRT with a dozen lines of Python; timestamps cost nothing extra.
- [What does TranscriptsDisabled mean, and can I still get the YouTube transcript?](https://okdata.online/answers/youtube-transcripts-disabled/): TranscriptsDisabled means the library found no caption track it could read for that video. Ask a second source to tell a truly captionless video from a blocked request: the YouTube transcript endpoint returns the text for $0.004, or not_found when the video really has no captions; either answer costs the same.
- [How do I build a RAG chatbot over YouTube video transcripts?](https://okdata.online/answers/youtube-transcripts-rag-chatbot/): Search YouTube for the videos you want ($0.004 per page of about 20), fetch each transcript with timestamps ($0.004 a video), cut the segments into time windows and embed them in any vector store. A 200-video knowledge base costs about $0.84 to collect, and every answer can link to the exact second it came from.
- [yt-dlp says "Sign in to confirm you're not a bot". How do I get YouTube subtitles without cookies?](https://okdata.online/answers/yt-dlp-not-a-bot-subtitles/): If you only need the words and the numbers, skip the download: the YouTube transcript endpoint returns the subtitles with timestamps for $0.004, and YouTube video detail returns title, description, tags, views and likes for $0.002. No cookies, no signed-in account, nothing runs from your server's IP.
