Chat conversationally with any YouTube video, generate chapter breakdowns, and extract key takeaways in seconds.
YouTube-Chat-AI is an open source productivity tool developed by Alex Santos. It ingests video URLs, fetches official or auto-generated transcripts, and leverages Gemini Flash 2.0 to answer detailed user inquiries.
From URL submission to transcript normalization, semantic chunking, and multi-turn dialogue.
Design choices, architectural patterns, and engineering decisions implemented for extreme reliability and developer experience.
Extracts subtitles and timed transcripts without downloading bulky video files, drastically reducing bandwidth and processing time.
Capitalizes on Gemini's massive context window and lightning-fast token generation to summarize and cite timestamps accurately.
Clean, responsive Python UI with streaming replies, markdown formatting, and exportable chat histories.
Pre-configured prompt templates generate comprehensive study guides and executive video briefs automatically.
Real-world usage, terminal configuration, and technical integration commands.
import gradio as gr
from youtube_transcript_api import YouTubeTranscriptApi
from google import genai
client = genai.Client()
def chat_with_video(video_url, user_question):
# Extracts video ID and captions
transcript = get_captions(video_url)
response = client.models.generate_content(
model="gemini-2.0-flash",
contents=f"Video Transcript: {transcript}
Question: {user_question}"
)
return response.text
gr.Interface(fn=chat_with_video, inputs=["text", "text"], outputs="text").launch()Key clarifications regarding licensing, compatibility, deployment, and security.
Yes, it automatically falls back to YouTube's auto-generated captions.
Thanks to Gemini Flash 2.0's million-token context window, videos exceeding 5 hours can be parsed easily.
Google AI Studio offers generous free-tier quotas for Gemini Flash models.
Yes, you can ask questions in English about a video recorded in any language.