Summarize YouTube video transcripts in Discord with Gemini and Supabase

YouTube Video Transcript Summarizer — Discord Bot

> Paste a YouTube URL into a Discord channel and this workflow automatically extracts the transcript, uses an LLM to generate a concise summary, and stores everything in a database — all in seconds.

> Self-hosted n8n only. This workflow uses the Execute Command node to run yt-dlp inside the n8n container. This requires shell access, which is only available on self-hosted instances (Docker, VPS, etc.) — it will not work on n8n Cloud.

Import this workflow into n8n

Prerequisites

Credentials

What It Does

When a user pastes a YouTube URL into a Discord channel, the workflow:

1. Detects the YouTube URL using RegEx (supports youtube.com, youtu.be, shorts, live)
2. Extracts the video's subtitles (English and Vietnamese) and metadata using yt-dlp
3. Cleans the raw VTT subtitle file into plain-text transcript
4. Summarizes the transcript using an LLM (Gemini 2.5 Flash) into a TLDR + detailed summary (in the original language)
5. Stores the video metadata, full transcript, and AI summary in a Supabase database
6. Logs every run (success or error) to a separate runs table for tracking
7. Chunks long summaries into Discord-safe messages (≤2000 characters each)
8. Replies in Discord with the video title, stats, and the full summary

Non-YouTube messages get a friendly "not a YouTube link" reply. Errors are caught, classified, logged to the database, and reported back to Discord.

How It Works

Main Flow (Happy Path)

Discord Trigger → Extract YouTube URL → Is YouTube URL?
 ├─ Yes → yt-dlp Get Metadata → Parse Metadata → Read Subtitle File → Parse Transcript
 │ → Message a model (Gemini) → Prepare Insert Data → Save to Supabase
 │ → Prepare Success Log → Log Run → Prepare Messages for Discord → Discord Reply
 └─ No → Discord Not YouTube Reply

Error Flow

Error Trigger → Prepare Error Data → Log Run Error → Discord Error Reply

Node Breakdown

Setup Guide

1. Discord Bot

1. Go to the Discord Developer Portal
2. Create a new Application → Bot
3. Enable Message Content Intent under Privileged Intents
4. Copy the Bot Token
5. Invite the bot to your server with Send Messages + Read Messages permissions
6. In n8n, create a Discord Bot Trigger credential (for listening) and a Discord Bot credential (for sending replies)
7. Update the guild ID and channel ID in the Discord Trigger node and all Discord reply nodes

2. yt-dlp

yt-dlp must be installed in your n8n container. For Docker-based installs:

docker exec -it n8n apk add --no-cache python3 py3-pip
docker exec -it n8n pip3 install yt-dlp

Optional: Place a cookies.txt file at /home/node/.n8n/cookies.txt to avoid age-gated or bot-detection issues.

3. Google Gemini API

1. Go to Google AI Studio
2. Click Create API Key and copy it
3. In n8n, click the Gemini node → Credential → Create New
4. Paste your API key and save

4. Supabase

1. Create a project at supabase.com
2. Go to Settings → API and copy the URL and anon key
3. In n8n, create a Supabase credential with your URL and API key
4. Run the SQL below in the Supabase SQL Editor to create the required tables

Supabase SQL

-- Videos table: stores video metadata, transcript, and AI summary
CREATE TABLE videos (
 video_id TEXT PRIMARY KEY,
 title TEXT,
 channel TEXT,
 upload_date TEXT,
 duration INT,
 view_count INT,
 description TEXT,
 transcript TEXT,
 ai_summary TEXT,
 thumbnail_url TEXT,
 channel_id TEXT,
 date_added TIMESTAMPTZ DEFAULT now()
);

-- Runs table: logs every workflow execution (success or error)
CREATE TABLE runs (
 video_id TEXT PRIMARY KEY,
 process_status TEXT NOT NULL,
 error_type TEXT,
 notes TEXT,
 date_added TIMESTAMPTZ DEFAULT now()
);