Hindi transcript is not returned correctly and language/API documentation is unclear

[shubhamkumarnayak]

Hi, thanks for maintaining this project.

I am trying to use the hosted YouTube Transcript API for Hindi YouTube videos, but Hindi transcript support is not working as expected. The API response also seems inconsistent with the claim that it supports multi-language transcripts.

API documentation shown by the hosted endpoint

The hosted endpoint returns this information:

{
  "description": "YouTube Transcript API by Jay Paun",
  "usage": "POST /transcript with JSON body { 'url': 'youtube_url' }",
  "notes": [
    "Supports multi-language transcripts including Hindi, English, and more.",
    "More powerful than most transcripters."
  ],
  "example request body": {
    "url": "https://www.youtube.com/watch?v=gTfsFeXgoKI"
  }
}

So I used the documented request format.

Request

POST https://youtube-transcript-api-tau-one.vercel.app/transcript
Content-Type: application/json

{
  "url": "https://www.youtube.com/watch?v=gTfsFeXgoKI"
}

Actual response

The API returns status: success, but the transcript is not a readable Hindi transcript:

{
  "status": "success",
  "url": "https://www.youtube.com/watch?v=gTfsFeXgoKI",
  "length": 55,
  "segments": 1,
  "transcript": "0:{\"a\":\"$@1\",\"f\":\"\",\"b\":\"i-dEv7N7_G2B1apWtEAsb\"} 1:null"
}

Expected response

Since the API says it supports multi-language transcripts including Hindi, I expected either:

1. A readable Hindi transcript, or
2. A clear error saying Hindi transcript is not available, along with available languages.

Example expected failure response:

{
  "status": "error",
  "message": "Hindi transcript is not available for this video",
  "available_languages": ["en", "hi"]
}

Problems noticed

1. The API says it supports Hindi, but there is no documented way to request Hindi.
2. There is no documented parameter like lang, language, or language_code.
3. The API returns status: success even though the transcript value is not usable text.
4. The response does not say which language was selected.
5. The response does not return available transcript languages.
6. The response does not explain whether it used manual captions, auto-generated captions, or a fallback language.
7. Some documentation/examples mention video_url, but the hosted API expects url.

Request

Please clarify and document:

1. Correct request body format.
2. How to request Hindi transcript.
3. Whether Hindi auto-generated captions are supported.
4. Whether a language parameter is available.
5. What fallback behavior is used when the requested language is not available.
6. Why the API returns status: success with unreadable transcript text.
7. Whether available languages can be returned in the response.

This would help developers reliably use the API for Hindi and other non-English transcripts.