How to Use Gemini 3 Analyze Video in n8n Automations

TL;DR

Gemini 3's analyze video feature lets you drop a video link into Gemini and get an accurate summary with timestamps, then interrogate the content by asking where specific topics were discussed. Integrated into n8n through a single Gemini analyze-video node, it can turn viral videos into custom scripts, Sora prompts, or auto-published LinkedIn and Twitter posts.

Transcript

This is probably Gemini Tree's most criminally underrated feature because it's useful for app development, creative agency work, viral content creation, and so much more. This is Gemini's analyze video feature, and in this lesson, I'll show you why it's so powerful, how to use it, and exactly where it fits to automate your creative work. Let's get ... Read More

Key Insights

• Gemini's analyze video feature can summarize a video from just its link with good accuracy, provide timestamps on key moments, and let you interrogate the video by asking where a specific topic, like scaling laws, was discussed so it points you to exactly where to listen.
• In n8n, the only node you need for this is the Gemini node set to the analyze video action, which replaces the traditional multi-step approach of using a paid scraper like Apify to grab the video first and then feeding it to a separate large language model across multiple nodes.
• Setting up the Gemini credential is a one-time step that requires an API key from Google AI Studio, obtained by clicking create API key, giving it a name, selecting the default Gemini project, and creating the key, after which all Gemini features work inside n8n.
• The recommended model at the time of recording is Gemini 3 Pro preview, which accepts a video URL and a text prompt, and can handle a long system guideline because it is Gemini, allowing detailed instructions on tone and output format.
• A too many requests error from Google is misleading because the actual fix is to set up billing on your API key through Google Cloud, which unlocks the free trial rather than indicating an overload problem.
• New Google users receive $300 in free credits, and running this analysis is cheap enough that 100 runs cost only about 10 to 20 cents in credits, though a card is required mainly to prevent bad actors from creating multiple free accounts.
• A Telegram agent hooked to n8n lets you share a viral YouTube short from your phone, and the analyze video node reviews it, understands the context, and generates a brand new script in a custom tone of voice such as Alex Hormozi's for instant recording.
• Because n8n is modular, the same node can send a long form video, over an hour, to Gemini which creates a LinkedIn or Twitter post that gets auto-published through Blotato, a tool that automates publishing across nine social media platforms including LinkedIn and Twitter.

Questions & Answers

Q: What is Gemini 3's analyze video feature and what can it do?

Gemini's analyze video feature lets you grab a video link, drop it into Gemini, and ask it to summarize the video, which it does with really good accuracy and even provides timestamps on key moments. You can also interrogate the video, so if you are curious where a specific topic like scaling laws was discussed, Gemini reads the transcript and pinpoints exactly where you need to listen. It is described as useful for app development, creative agency work, and viral content creation.

Q: How do you set up the Gemini analyze video node in n8n?

In n8n, find the Gemini node and select the analyze video action to add it to your workflow. The only setup step is creating a Gemini credential, which you do once. Create a new credential, which requires an API key obtained from Google AI Studio by clicking create API key, giving it a name, and selecting a project. Google creates a default Gemini project if you have none. Copy the key, paste it, save, and you can use Gemini models within n8n or other automation tools.

Q: Which Gemini model should you use for analyzing video?

The video recommends using Google's latest model, which at the time of recording is Gemini 3 Pro preview. You select this model in the node after setting up your credential. In the node you can give Gemini a prompt and paste in a URL of any video to analyze. Because it is Gemini, you can provide a pretty long system guideline with detailed directions on how to analyze the video and what output format you want, such as producing only a script and nothing else.

Q: Why does the too many requests error appear and how do you fix it?

When you first execute the analyze video step, you may get an error from n8n or Google about the service receiving too many requests. The video explains this error is quite misleading because the actual way to alleviate it is to set up your free trial over at Google AI Studio. You go back to where you created your API key and set up billing by configuring a Google Cloud billing account. Once billing is set up and you re-execute the step, it gives you an analysis of the video.

Q: How much does it cost to run Gemini video analysis?

The cost of running this is described as super cheap. New users get $300 in free credits from Google right off the bat. Running the analysis 100 times costs only something like 10 to 20 cents in credits from Google's side, which the video calls very generous. Setting up billing requires adding your card, but the video notes Google mostly does this to prevent bad actors from creating multiple free accounts on their site, not because the service is expensive to use.

Q: How can you turn a viral video into a custom script using n8n and Telegram?

If you are a content creator in an educational niche browsing YouTube shorts and find one going viral, you can hit share and send it to a Telegram agent hooked up to a simple n8n automation. The analyze video node reviews the video, understands the context, and generates a brand new script you can customize to your style or tone of voice. In the example it generated a script in Alex Hormozi's tone. You can then record it on your phone or extend the workflow to send the script to a tool like HeyGen to create an AI avatar video.

Q: How does the analyze video node compare to the traditional scraping method?

Traditionally, to do something like this you needed to use a paid scraper like Apify to grab the video first, then feed it to a large language model powered by OpenAI or Gemini, and set up multiple nodes to get the result. The analyze video node changes this because a single Gemini node does what previously required a scraper plus a separate LLM across multiple nodes. This makes setup much simpler, though the basic example does not yet automatically create finished content or publish to social platforms without additional modular nodes.

Q: How can you auto-publish social media posts from a long video with this workflow?

Because n8n is modular and flexible, you can use the analyze video node for auto posting. In the example, you send a YouTube long form video, more than an hour long such as a Dwarkesh podcast episode that is an hour and 36 minutes, by sharing the link to the Telegram agent. Gemini analyzes the video and creates a LinkedIn or Twitter post, which is then auto-published by a tool called Blotato. Blotato lets you automate publishing for nine social media platforms, including LinkedIn and Twitter, and the demo resulted in a successful post.

Summary & Key Takeaways

• Jay introduces Gemini 3's analyze video feature as a criminally underrated capability useful for app development, creative agency work, and viral content creation. Gemini can summarize a video from its link with good accuracy, provide timestamps on key moments, and let you pinpoint where specific topics were discussed by reading the transcript.

• The feature becomes powerful inside n8n, a no code automation tool similar to Zapier or make.com, through a single Gemini analyze video node. Setup requires a one-time credential using an API key from Google AI Studio and selecting the Gemini 3 Pro preview model. A too many requests error is fixed by enabling billing.

• Practical automations include a Telegram agent that turns viral shorts into custom scripts in a chosen tone of voice, creating Sora video variations for niches like pets, and analyzing hour-long podcasts to auto-publish LinkedIn or Twitter posts via Blotato, which supports nine social platforms including LinkedIn and Twitter.