Turn long videos into clear text descriptions without watching them. Built for marketing and content teams that need fast tags, summaries, and compliance checks at scale. Helpful for social clips, ad reviews, and media libraries.

The run starts with a manual test. A Set node stores the video URL. An HTTP Request node downloads the file as binary. Another HTTP Request uploads that binary to Google Gemini using the upload endpoint with required headers, which returns a file link. A Wait step gives the upload time to finish. The next call uses the Gemini generate content endpoint with the file link and a clear prompt. A final Set node extracts the first candidate text into a field named videoDescription for easy use in later steps.

Use a Gemini API key through an environment variable named GeminiKey or with credentials. Expect review time to drop from many minutes to seconds per video while keeping wording consistent across assets. Common uses include content moderation notes, SEO friendly blurbs, and accessibility helper text. You can adjust the prompt to focus on brand safety, product scenes, speaker changes, or on screen text. The output is ready for tagging, catalog search, or sending to a CMS in a follow up workflow.

• Manual test run to control when analysis starts
• Set node captures the video URL as a simple input
• HTTP Request downloads the video as binary data ready for upload
• HTTP Request uploads the video to Google Gemini with required headers and content length
• Wait step pauses to ensure the upload is ready before analysis
• HTTP Request calls Gemini generate content with fileUri and a customizable prompt
• Set node maps the AI answer to videoDescription for clean handoff

• Reduce manual viewing time from 30 minutes to 2 minutes per video
• Automate up to 90 percent of metadata writing for videos
• Improve consistency of descriptions and tags by 80 percent
• Handle 5 to 10 times more videos without adding staff
• Connect AI results to downstream systems with one output field

1. Import the template into n8n: Create a new workflow in n8n > Click the three dots menu > Select 'Import from File' > Choose the downloaded JSON file.
2. You'll need accounts with Google Gemini. See the Tools Required section above for links to create accounts with these services.
3. In Google AI Studio, create an API key for the Gemini API. Copy the key and keep it secure.
4. In n8n Cloud, open Settings > Environment Variables and add a variable named GeminiKey with your API key. Save and reload the editor.
5. Open the Set Input node and replace the sample video_url with the URL of a video you have rights to use.
6. Open the Download video HTTP Request node and confirm it uses the video_url field and returns binary data in the data property.
7. Open the Upload video Gemini HTTP Request node and confirm the URL includes the key from {{$vars.GeminiKey}}, content type is binary data, input field is data, and the X Goog upload headers are present.
8. Check the Wait node and keep a short pause so the upload can complete. If needed, increase the delay for larger files.
9. Open the Analyze video Gemini HTTP Request node and confirm it calls the generate content endpoint, sends JSON, and uses the fileUri from the upload step. Edit the prompt to match your use case.
10. Open the Get Result Set node and confirm it maps candidates[0].content.parts[0].text to videoDescription.
11. Click Test workflow. Verify the Download step shows binary data, the Upload step returns a file.uri, the Analyze step returns candidates, and the final output includes videoDescription.
12. If you see 401 or 403 errors, check GeminiKey. If analysis fails, confirm file size and headers. If the output is empty, extend the Wait step or verify the fileUri value.