Transcription from YouTube video

Hello!

I am ready to offer an effective solution to your task using the Make and Dumpling AI integration — this will allow you to fully automate the process of transcribing YouTube videos into text.
🔧 How the solution works:
You add a link to a YouTube video (or the system automatically retrieves them).
Make sends the link to Dumpling AI.
Dumpling AI automatically extracts the audio and converts it into text.
The result is saved in Google Sheets, Airtable, etc.
💸 Terms and prices:
… Make: 1,000 operations per month on the free plan — enough for testing and pilot work.
Dumpling AI: 250 credits upon registration for free — enough for 125 videos (1 video = 2 credits).
🧾 Pricing after free limits:
✅ Make:
Core plan: $9/month — 10,000 operations
Pro: $16/month — 40,000 operations
✅ Dumpling AI:
Starter: $40/month — 120,000 credits per year
Additional credits: $10 for 1,000 (no expiration date)

I am ready to discuss details, adapt the solution to your business processes, and show a demo. I also have previous experience working on a similar task.

Show original

Translate

Hello!

I am interested in your project. I have extensive experience in automation and user action emulation (JavaScript, Selenium, Playwright), asynchronous and multithreaded parsing (Requests, WebSockets, HTTPX, BS4), data processing (Openpyxl, JSON, MySQL, MongoDB), as well as developing Telegram bots of various complexity levels (Telethon, Pyrogram, Aiogram).

I have worked with Whisper and video transcription with timestamps and translation using DeeplAPI. The best option would be to deploy everything on a server with at least 4 CPU / 8–12 GB RAM for transcription with the best preset.
We can adapt the original text with timestamps to meet the required specifications (for example, if third-party services will be used for processing or editing).

Contact me to discuss the details and deadlines of the project!

Show original

Translate

Hello, Petr
I have a solution I made for myself (macOS). It is implemented with bash scripts
It works according to the following algorithm - a video in .mp4 format is placed in the folder with the main script, the main script is launched, and an automaton is started under the hood
1. split: video + sound (original)
2. split: voice + background noises (music, noise, etc.)
3. converting voice to text (in the required language --> OpenAI)
4. reverse process: text to voice (in the required language --> OpenAI)
5. the result (.mp3 file) is placed in a folder next to the video
If desired, I can demonstrate this via video call
… It is also possible to work directly with YouTube (fetching the audio immediately and then converting as needed)
In the current version, I will give the script for a symbolic price of 6000 UAH (price for the time spent on writing)
For other questions - refinement, adaptation, reworking, etc., at the rate I indicated in the rate.
Maybe I can be useful...

Show original

Translate

I built a system that automatically translated video lessons from English to Ukrainian in a fully automatic mode. Used Make, but N8N is also possible. Feel free to contact me, I will be happy to help.

Show original

Translate

Hello!
The cheapest option will be your own PC with Whisper API installed, and then the process on the cloud.

Or a server from Hetser, calculate the power you need to cover, and then you can organize everything on it.

The stack will look like this
n8n + postgres + Whisper + (admin panel if needed, where you need it in Google Sheets)

Show original

Translate

Good afternoon.
Ready to take on your project.
I can develop automation for transcribing YouTube videos for you using no-code/low-code tools.
Write to me privately, we will discuss all the details and find the best solution for you.

Show original

Translate

Good afternoon. I am ready to write such a program. Let's get in touch and discuss the details

Show original

Translate

I made a similar project for myself. I downloaded videos from YouTube using yt-dlp, extracted the audio track, and sent it to Gemini Flash 2.0 for transcription and subtitle creation. But I had to tinker because Flash 2 incorrectly determined the length of the audio track and generated incorrect timestamps without a lengthy prompt.

Show original

Translate

Hello Petro! I understand your task: automatic extraction of text from YouTube videos, focusing on cost-effective efficiency and considering all expenses, including servers.

My experience and relevance
I have already implemented similar tasks, where it was necessary to convert audio into text and automate data processing. For example, in the project "AI Call Record Analysis Automation using Make," I created a prototype system that monitors audio files, transcribes them using Whisper (OpenAI), and then analyzes the text.
In addition, I have identical experience in the work you mentioned: "Content repurposing system for YouTube videos." In this project, I built an AI system using Airtable and Dumpling that transcribes, analyzes content, and generates fragments for various platforms (email, Telegram, Facebook, Instagram). This task involved extracting text from videos for further processing, which is a direct parallel to your current need.

This connection will help implement your project.
What specific sources of YouTube videos do you plan to use (particular channels, playlists, or random videos by search)?

Show original

Translate

The most productive and simple way is just to extract subtitles from the video. They come as a separate file. Just download the file. But there may be difficulties, as YouTube does not allow downloading all videos, depending on the license...

Or as an alternative, run an optimized mini AI model (there are many free and paid ready-made models) and read the video's audio track.
Obviously, processing will cost significantly more.

In any case, it is important to combine this process with saving the video itself to avoid repeated unnecessary operations.

Show original

Translate

Hello,

I can create a solution for you that will extract video lists from your links in the Google Spreadsheet document or from your database and prepare transcriptions for you in the required format. It is even possible to do this with speaker highlighting (suitable for further processing) or in the standard .srt format.

Example of JSON output with speakers:

{
"lines" : [
{
… "endTime" : "00:00:55,460",
"speakerDesignation" : "Speaker 1",
"startTime" : "00:00:52,659",
"text" : "text"
},
{
"endTime" : "00:01:10,800",
"speakerDesignation" : "Speaker 2",
"startTime" : "00:01:06,300",
"text" : "I'm also speaking"
},
...

It can be set up so that after adding a video to the database, you receive it automatically via email or in Telegram.

Show original

Translate

Hello! I have a budget solution proposal for this task. Write in private messages, let's discuss.

Show original

Translate