AI product photo to video generator

What it is

This n8n workflow takes one product photo and a sentence about the video you want, and hands back a short commercial-style clip. Eight nodes run one after another. A form takes the photo, Cloudinary stores it at a public address, and Gemini 2.5 Pro writes a YAML description of the product with its brand and its colors as hex codes. An agent named Creative Director turns that description and your sentence into a JSON shot plan. A sub-workflow you supply sends the plan and the photo to Veo 3.1 Fast on Kie.ai, and the last node downloads the clip.

Who it suits

Shop owners and small brands that want a moving ad for every product without booking a shoot, and the marketers and agencies who make those ads for them.

You need some n8n: importing a file, adding credentials, building one small sub-workflow. If that is new to you and you would rather not learn it, skip this one. Its author puts setup at about 11 minutes, which assumes the four accounts already exist.

Where to use it

Any n8n you run will do, as long as it can reach Cloudinary, Google, OpenRouter and Kie.ai.

• n8n Cloud: import the file into a new workflow there
• Your own n8n server: the same file, the same steps
• Ad and social channels: the clip comes back as a file, ready for an Instagram ad, a Reel or a Short
• A drive or another app: add a node after Download Video to save the clip or send it on

Setup

Get it saves a zip with workflow.json and Mindket’s setup notes in your downloads folder. Have these ready before you import it.

1. In n8n, create a workflow, open its menu, choose Import from File and pick workflow.json.
2. Open Upload *Photo1 and put your cloud name in its address where it says <YOUR_CLOUD_NAME>.
3. In the same node, swap YOUR_PRESET_NAME for your preset and YOUR_FOLDER_NAME for a folder.
4. Add a Google Gemini (PaLM) API credential and select it in Analyze an image.
5. Add an OpenRouter credential and select it in OpenRouter Chat Model.
6. Open the red node after Creative Director, the call to the Kie.ai sub-workflow, and pick your own video sub-workflow.
7. Save, send a test photo through the form’s test address, then press Publish.

The sub-workflow is not in the file. Its node pointed at a workflow in its author’s own n8n, so it arrives empty and marked red. Yours receives these four fields and returns the clip’s address as resultUrls:

{
  "video_prompt": "<the shot plan, as JSON text>",
  "aspect_ratio": "16:9",
  "imageUrls": ["<the photo’s Cloudinary address>"],
  "generationType": "REFERENCE_2_VIDEO"
}

Kie.ai’s Veo pages, read on 2 November 2026, describe the two calls inside it. Post those fields to the generate endpoint with your key as a Bearer token, then ask record-info for the task until it reports success. The resultUrls field comes back as a JSON string, so parse it before you return it.

POST https://api.kie.ai/api/v1/veo/generate
GET  https://api.kie.ai/api/v1/veo/record-info?taskId=<taskId>

A signed preset works too: add an HTTP Basic Auth credential with your Cloudinary API key and secret, and select it under the upload node’s Authentication.

How to use it

Open the form from On form submission, attach a product photo and write the video you want in Description, in a sentence. The form confirms at once. The run goes on behind it: upload, analysis, shot plan, render, download. The clip lands in the run’s execution data as a file, so open the run under Executions to see it.

Example

A photo of a bottle of cold brew coffee, with this in Description:

Ice cubes drop into a glass beside the bottle on a sunny cafe counter, slow motion, warm morning light

Gemini’s YAML names the product type, the brand if the label shows one and the bottle’s colors as hex codes. The Creative Director answers in the parser’s shape, trimmed here. It leaves the bottle’s look out of the plan, because its prompt says Veo takes that from the photo.

{
  "scene": "…",
  "style": "…",
  "camera": { "type": "…", "angle": "…", "lens": "…", "movement": "…" },
  "lighting": { "type": "…", "mood": "…", "time": "…" },
  "audio": { "music": "…", "effects": ["…"] },
  "motion": { "pace": "…", "style": "…", "transitions": "…" },
  "technical": { "resolution": "…", "duration": "…", "color_palette": "…", "effects": "…" },
  "narrative": { "beginning": "…", "middle": "…", "end": "…" }
}

Kie.ai renders the clip from the plan and the photo, and Download Video fetches it from resultUrls.

For newcomers and experts

If you are new. n8n draws a workflow as boxes joined by lines, and each box, a node, does one job. A credential is a key n8n keeps for you and hands to the node you pick it in. Run the form from its test address first. Every node then shows what it received and what it sent, which is where a wrong preset or key turns up.

For experts. Three models share the work, and each step names the one it uses.

You can point OpenRouter at another model too. The agent carries the YAML’s color_palette_hex into the plan’s color_palette, so a brand’s colors reach the scene.

Good to know

Each run costs money, and the rest of the risks come from the open form and the public photo.

• Billing: Google, OpenRouter and Kie.ai each bill their own part of a run
• Open form: once the workflow is published, anyone with the form’s address can start a run on your keys, so turn on the form trigger’s basic authentication or keep the address to yourself
• The photo: an unsigned upload makes it public on Cloudinary, and it also goes to Google, OpenRouter and Kie.ai
• File types: the form’s file field takes any file type, so send photos only
• Credentials: the JSON holds none, and nothing from its author’s account