YouTube

YouTube Transcript

Fetch a video's transcript as plain text, or as timed SRT subtitles with start and end times — one synchronous call, as JSON. Videos without captions are transcribed from their audio. Costs 5 credits per call.

POST/api/v1/youtube/transcript

Authorizations

Authorizationstringheaderrequired

Bearer authentication header of the form Bearer <token>, where <token> is your Scavio API key (e.g. Bearer sk_live_your_key).

Body

application/json

video_idstringrequired

The YouTube video id (11 characters) or any video URL: watch, Shorts (youtube.com/shorts/...), youtu.be and embed links all work.

Example: n61ULEU7CO0

languagestringdefault:en

Caption language code to fetch (e.g. en, es, fr). If the video has no track in that language, content is null and the response lists available_languages so you can retry with one of them. The call is still billed.

formatenum<string>default:text

Output format: text for a plain transcript, srt for timed subtitles. Videos without captions are transcribed from their audio, which takes longer (about 10 to 15 seconds). For a video with no speech, an srt request returns content null with the message "No speech detected in this video."

Request

from scavio import ScavioClient

client = ScavioClient(api_key="sk_live_your_key")
results = client.youtube.transcript(video_id="n61ULEU7CO0", format="srt")
print(results)

Response

"video_id": "n61ULEU7CO0",

"language_code": "en",

"language_name": "English",

"format": "txt",

"content": "Welcome back to the channel. Today we are going to walk through building a retrieval pipeline from scratch..."

"response_time": 1633,

"credits_used": 5,

"credits_remaining": 4747

}