anaksor-ocr
Sokheng · Kit
Reads Khmer and English text from documents, receipts and photos of signs, using a detector for text lines and a recogniser for each line. Suits developers who…
An n8n workflow that has Gemini describe a video from a file you upload or a YouTube link you paste. It waits for Google to finish with the file, then posts the written description to a converter address you set, which can turn it into a Google Doc.
Gemini video description from a file or a YouTube link is an n8n workflow in the Productivity category, listed by Mindket. It is free. This is version 2, last updated on . Download it on this page or copy its JSON, then import it into an n8n canvas. Nothing runs until you activate it in n8n.
File or link form
It starts when someone sends one of two Insert Video forms. The first takes a file, the second a link:
A detailed written description of the video, ready to become a Google Doc
What you need before you start.
Video analysis
Takes the uploaded file or the YouTube address and writes the description with Gemini 2.0 Flash.
NeedsGemini API key, as an HTTP Query Auth credential
Document
Receives the description through the converter address you set, with your Google sign-in.
NeedsGoogle Docs OAuth2 credential, and a converter address
Checked by Mindket on 13 Nov 2026
This n8n workflow has Gemini describe a video for you, from a file you upload or a YouTube link you paste. Two forms start it, one for each. A file goes to Google’s file store, the workflow waits, and Gemini 2.0 Flash then writes a long description of the whole clip. A link skips the upload and goes straight to the same model with a request for a detailed summary. Either way the text ends in a request to a converter address that you supply, which can turn it into a Google Doc.
Editors, researchers and content teams who need a written account of a video without watching all of it. Skip it if you want a finished Google Doc with no extra work. The converter step is yours to supply, and the file does not include one.
| Industry | Who uses it | How they use it | Example |
|---|---|---|---|
| Video production | Editors and archivists | Upload raw footage and keep the description beside the file name | An editor uploads a screen recording and gets a written account of each step to search later. |
| Accessibility | Teams that caption or describe video | Use the description as a first draft of what happens on screen | A team drafts notes on the visuals of a training video, then corrects them by hand. |
| Research and media | Analysts and journalists | Paste the link of a public video and read the summary before deciding to watch | A journalist pastes a conference talk’s link and gets a detailed summary. |
| Marketing | Content and social teams | Summarize videos from partners or rival channels | A social team summarizes a rival’s product demo from its link. |
| Education | Teachers and course makers | Turn a recorded lesson into the base for study notes | A teacher uploads a lesson recording and edits the description into notes. |
Any n8n that can reach Google’s Gemini API will run it. The two forms give you two web addresses to share with whoever has videos to send.
Get it saves a zip with workflow.json and Mindket’s setup notes in your downloads folder. Its author puts setup at about 10 minutes. The original copy sent the text, with your Google sign-in, to a converter on a private domain that is neither yours nor Mindket’s. This copy leaves that address blank, so you decide where your text and sign-in go.
| Need | Detail |
|---|---|
| n8n | On n8n Cloud or on a server of your own |
| Google AI Studio | A key for the Gemini API |
| Google Docs | A Google Docs OAuth2 credential, for your converter or for Google Docs nodes |
| A converter | A web address that takes the text and files it as a Google Doc, or two Google Docs nodes in its place |
workflow.json.key and your Gemini key as the value. Select it in Upload File, Get Analysis and YouTube Video.HTTP Request and HTTP Request1.https://YOUR_MARKDOWN_TO_DOC_ADDRESS with your converter’s address, or swap them for Google Docs nodes.Each of the two requests posts one field, output, with the description in it, and n8n adds your Google Docs sign-in. What your converter does with them is up to you.
{
"output": "<the description Gemini wrote>"
}To skip the converter, swap each request node for two Google Docs nodes: Create, then Update with Insert, which writes the text into the new document.
Open the form you need from its trigger, On form submission for a file or On form submission1 for a link. Both forms carry the title Insert Video, so label each address when you share it. The file form takes one required video. The link form takes one required address. Each form confirms at once, and the run goes on behind it.
| To change | Where |
|---|---|
| What Gemini is asked | The text in the jsonBody of Get Analysis for files and of YouTube Video for links |
| Model | gemini-2.0-flash in the address of both calls |
| Waiting time | Wait before the first check and 5 seconds between retries |
| Where the text goes | The two request nodes after Video Analysis and Get Results |
A team lead uploads a short .mp4 of someone building a pivot table. Gemini describes the clip from start to end, Video Analysis keeps that text in one field, and the field travels to your converter as output.
{ "Video Analysis": "<Gemini’s description of the whole clip>" }If you are new. n8n draws a workflow as boxes joined by lines, and each box, a node, does one job. This file has 19, but eight are sticky notes that only label the stretches. Gemini is Google’s AI model, and the key you add is how Google knows the calls are yours. Send a short test video first. Each node then shows what it received and sent.
For experts. Both branches use one model, gemini-2.0-flash, through Google’s Generative Language API with the key in the query string.
| Node | Endpoint | Sends | Gets back |
|---|---|---|---|
Upload File | /upload/v1beta/files | The video as binary data, with headers that start and finish the upload in one call | A file record with its uri and mimeType |
Get Analysis | gemini-2.0-flash:generateContent | The fileUri, the mimeType and a request for a full description | The text at candidates[0].content.parts[0].text |
YouTube Video | gemini-2.0-flash:generateContent | The link as file_data.file_uri and a request for a detailed summary | The same field |
Get Analysis has its error output switched on. If the request fails, for example because Google is still processing the file, the error line leads to 5 seconds. That node sends the call round again. Wait and 5 seconds have no settings of their own, so both use n8n’s default of five seconds for this version of the node. The link branch has no wait and no retry.
A wrong key or a blocked file loops until you stop it, and the converter receives your Google sign-in.
An n8n workflow. Send a video file or a YouTube link through one of its two forms, and Gemini writes a long description
of the video. The workflow then posts it to a converter address that you supply.
workflow.json: the workflow as n8n exports it, 19 nodes. Eight of them are sticky notes that label the stretches. Ithttps://YOUR_MARKDOWN_TO_DOC_ADDRESS. The setup note on the canvas also lost its author line.README.md: these notes.workflow.json.key and your Gemini key as the value. Select it in the nodeUpload File (its name ends with a space), in Get Analysis and in YouTube Video.HTTP Request and HTTP Request1.https://YOUR_MARKDOWN_TO_DOC_ADDRESS with your converter's address.On form submission takes a file, On form submission1 takes a link) and sendEach of the two requests posts one field, output, holding the description, and n8n adds your Google Docs sign-in to the
request. Nothing in the file reads the reply. Point the address only at a service you trust, since it receives that
sign-in and the whole description.
To skip a converter, delete HTTP Request and HTTP Request1. In their place, join a Google Docs node set to Create,
and after it one set to Update with the Insert action, to Video Analysis and to Get Results. Point the Update node at the new
document's ID from the Create node.
jsonBody of Get Analysis (files) and YouTube Video (links).gemini-2.0-flash in the address of both calls.Wait pauses before the first check and 5 seconds between retries. Neither has a setting of its own, soGet Analysis goes through 5 seconds back to the call,README.md 3 KB workflow.json 13 KB
Updated by Mindket to the current version of this kit.
Sokheng · Kit
Reads Khmer and English text from documents, receipts and photos of signs, using a detector for text lines and a recogniser for each line. Suits developers who…
Qwen · Kit
Qwen2.5-7B-Instruct as GGUF files in eight quantization sizes, to chat with locally through llama.cpp. For people who want a capable 7.6B model on their own co…
ACE-Step · Kit
Generates music from text with a 4B diffusion decoder, 50 steps and guidance control, aiming for higher audio quality than the base. For creators with a GPU of…
op7418 · Skill
Skill that makes single-file HTML slide decks, slide illustrations and platform covers in two visual styles, with a presenter mode for rehearsal. For speakers…
zarazhangrui · Skill
Builds HTML presentations from scratch or from a PowerPoint file and shows visual style previews to pick from, for people who want web slides with no CSS or Ja…
iptv-org · Kit
Collects publicly available live TV channels from many countries into playlists that any video player with live streaming can open. Made for viewers and for de…