Build an image-to-video ad workflow with n8n, Nano Banana and Veo3

Set up the host and the workflow inputs

The September 2025 tutorial uses n8n on a Hostinger VPS to orchestrate product ads. Telegram supplies an entry point, Google Sheets and Drive support data handling, Nano Banana generates images through OpenRouter and Veo3 produces video through fal.ai. These are separate services with separate inputs and results.

1. Open https://hostinger.com/AILUKE and follow the hosting setup from 0:28. Choose the server through the demonstrated requirements and inspect the imported workflow at 4:42 before adding credentials.
2. Define one permitted product reference, the ad’s message and the expected output. Keep the template’s trigger, data record and image-to-video stages identifiable so a failure can be traced to its actual step.

Connect Telegram and check the saved record

The Telegram connection begins at 8:12 and Sheets logging at 12:21. A received message and a saved record are different checkpoints. The source promises a community template but does not supply a direct file URL, so the workflow structure in the recording remains important for following the connections.

1. Follow the demonstrated bot and credential setup, then send a small test input. Verify which workflow received it and compare the resulting record with the supplied product and brief.
2. Open the saved sheet or file destination and inspect fields, links and missing data. Confirm that the next generation step reads that same record rather than a leftover sample or unrelated input.

Review the image before spending on video

The sequence improves prompts at 17:27, generates product images at 20:35 and video at 24:31. The description’s less-than-$0.50 claim is historical, not a current price guarantee. Repeated failed image or video calls can change the cost of a usable result.

1. Compare the generated image with the product reference, checking identity, label and framing. Correct a specific mismatch before passing the image to the Veo3 stage.
2. Watch the final video, locate its output file and check product continuity, motion and text. Record usage and failures alongside the result, then test another small brief before expanding the automated workload.

The takeaway

Treat hosting, trigger delivery, stored data, image fidelity and video output as separate checks. The useful automation delivers a reviewed asset from the intended input, not merely a successful workflow status.