"""
title: YouTube Transcript Fetcher
author: RVCKVS
description: Fetches YouTube video transcripts using the video URL without needing an API key.
required_open_webui_version: 0.4.0
requirements: requests
version: 1.0.0
licence: MIT
"""
import requests
import re
import xml.etree.ElementTree as ET
from pydantic import BaseModel, Field
class Tools:
class UserValves(BaseModel):
"""
User-configurable parameters.
"""
language: str = Field(
"en", description="Preferred transcript language (default: en)"
)
def __init__(self):
"""Initialize the tool with user-configurable valves."""
self.user_valves = self.UserValves()
async def fetch_youtube_transcript(
self, video_url: str, __event_emitter__=None
) -> str:
"""
Fetches the transcript of a YouTube video given its URL without an API key.
:param video_url: The URL of the YouTube video
:return: The video transcript as a string
"""
# Notify user about the extraction process
await __event_emitter__(
{
"type": "status",
"data": {"description": "Extracting video ID...", "done": False},
}
)
video_id = self.extract_video_id(video_url)
if not video_id:
return "Invalid YouTube URL"
await __event_emitter__(
{
"type": "status",
"data": {"description": "Fetching video page...", "done": False},
}
)
try:
response = requests.get(f"https://www.youtube.com/watch?v={video_id}")
response.raise_for_status()
except requests.RequestException as e:
await __event_emitter__(
{
"type": "status",
"data": {"description": "Failed to fetch video", "done": True},
}
)
return f"Failed to fetch YouTube video page: {e}"
body = response.text
await __event_emitter__(
{
"type": "status",
"data": {"description": "Extracting captions data...", "done": False},
}
)
start_index = body.find('"captionTracks":')
if start_index == -1:
await __event_emitter__(
{
"type": "status",
"data": {"description": "No captions available", "done": True},
}
)
return "No captions found for video"
end_index = body.find("]", start_index)
if end_index == -1:
return "Invalid captions data"
captions_data = body[start_index : end_index + 1]
caption_url = self.extract_caption_url(captions_data)
if not caption_url:
await __event_emitter__(
{
"type": "status",
"data": {"description": "No English captions found", "done": True},
}
)
return "No suitable captions found"
await __event_emitter__(
{
"type": "status",
"data": {"description": "Downloading transcript...", "done": False},
}
)
try:
caption_response = requests.get(caption_url)
caption_response.raise_for_status()
except requests.RequestException as e:
return f"Failed to fetch transcript: {e}"
transcript = self.extract_transcript_from_xml(caption_response.text)
if not transcript:
return "Transcript extraction failed"
await __event_emitter__(
{
"type": "status",
"data": {
"description": "Transcript fetched successfully!",
"done": True,
},
}
)
# Send the transcript back to the LLM for further processing
await __event_emitter__(
{
"type": "message",
"data": {"content": f"Transcript has been fetched!\n\n"},
}
)
# Returning the transcript so the LLM can also process it internally
return f"Here's the transcript, : {transcript}.\n\n"
def extract_transcript_from_xml(self, xml_content: str) -> str:
"""
Extracts readable text from YouTube XML captions.
:param xml_content: The XML content of the captions
:return: The cleaned transcript text
"""
try:
root = ET.fromstring(xml_content)
transcript_lines = [
text.text.replace("'", "'").replace("&", "&")
for text in root.findall(".//text")
]
return " ".join(transcript_lines)
except ET.ParseError:
return ""
def extract_video_id(self, url: str) -> str:
"""
Extract YouTube video ID from URL.
:param url: The YouTube video URL
:return: Extracted video ID or empty string if invalid
"""
match = re.search(r"v=([a-zA-Z0-9_-]{11})", url)
return match.group(1) if match else ""
def extract_caption_url(self, captions_data: str) -> str:
"""
Extracts the caption URL from captions metadata.
:param captions_data: The caption metadata string
:return: Extracted caption URL or empty string
"""
url_start = captions_data.find('"baseUrl":"')
if url_start == -1:
return ""
url_start += len('"baseUrl":"')
url_end = captions_data.find('"', url_start)
if url_end == -1:
return ""
return captions_data[url_start:url_end].replace("\\u0026", "&")
