YouTube video summarized from its transcript through a form

YouTube video summarized from its transcript through a form

An n8n workflow with a web form. You paste a YouTube link, an Apify actor fetches the video's transcript, and an OpenAI
model writes a summary of it. The summary sits in the run's output in n8n.

What is inside

• workflow.json: the workflow as n8n exports it, 5 nodes beside 5 sticky notes. It holds no credential.
• README.md: these notes.

What we changed from the original copy

Mindket filled in the gaps the copy left:

• The request node's address read "Apify API_KEY Here ???", a note to fill in. It now holds the address of Apify's
  run-and-return endpoint for the actor topaz_sharingan~Youtube-Transcript-Scraper, with <YOUR_APIFY_TOKEN> after
  token=. The request body already matched that actor's input, a list called startUrls. Mindket read that in the Apify
  store on 12 November 2026.
• The form's path was ddd. It is now youtube-summary.
• The chat model named no model, so n8n would fill in gpt-5-mini, which OpenAI shuts down on 11 December 2026. Mindket's
  copy names gpt-5.6-terra, the replacement OpenAI gives for it.
• One sticky note sent you to a page Mindket does not have. Mindket reworded it.
• No credential is inside.

Before you start

• An n8n instance, on n8n Cloud or your own server.
• An Apify account and its API token. Apify bills each run, so read the actor's page before you start.
• An OpenAI API key.

Set it up

1. In n8n, create a workflow, open its menu and choose Import from File. Pick workflow.json.
2. Open Request YouTube Transcript and replace <YOUR_APIFY_TOKEN> in the address with your Apify API token. To use
   another actor, swap its id into the address. Its input must take startUrls, or you change the request body.
3. Add an OpenAI credential and select it in Summarization Engine.
4. Save. Open YouTube video URL, use its test form address, paste a YouTube link into Full URL and submit.
5. Open the run in n8n and read the summary in the output of Summarization of a YouTube script.

How a run goes

1. YouTube video URL shows a form with one field, Full URL. A run starts when someone submits it.
2. Request YouTube Transcript posts the link to Apify as {"startUrls": ["<the link>"]} and waits for the actor to
   finish. Apify's docs, read on 12 November 2026, say this endpoint answers with a 408 when a run passes 300 seconds.
3. Summarization of a YouTube script takes each item the actor returns, uses all of its JSON as the text and cuts it into
   chunks of 1000 characters that overlap by 200. It then summarizes with the Map Reduce method and the node's default
   prompts.
4. Summarization Engine is the chat model it uses.
5. No Operation, do nothing ends the run.

Change the look

• The video source: the address and body in Request YouTube Transcript. Any service that returns a transcript works.
• The style of the summary: the method and prompts under the options of Summarization of a YouTube script.
• The model: Summarization Engine.
• The input: swap the form for a webhook or any trigger. The request body reads the field Full URL.
• Where the summary goes: add nodes after No Operation, do nothing, such as an email or a spreadsheet row.

Good to know

• The form answers "Your response has been recorded", the default text of n8n's form. It does not show the summary, which
  you read in the run's output.
• The form has no sign-in. Once the workflow is published, anyone with its address can start a run on your Apify and
  OpenAI accounts.
• A long transcript means many model calls, since Map Reduce summarizes each chunk and then joins the results.
• The video link goes to Apify, and the transcript goes to OpenAI.
• The actor's page says which videos it can read and what it returns. Check it before you rely on the summary.
• The JSON holds no credential, and the token in the address is a placeholder.