Character selfie vlog video from a one-line idea

What it is

This n8n workflow turns one line, such as a yeti reviewing a hiking trail, into a short video shot as a selfie. The character holds the phone, delivers a dry line or two, and looks like an influencer filming a vlog. Eight nodes run in a line. An agent on gpt-4.1 writes the Veo 3 prompt together with a title and a social caption. Kie.ai renders it with Veo 3 Fast. Two minutes later the workflow downloads the MP4.

1. A form takes a one-line idea
2. gpt-4.1 writes the prompt, a title and a caption
3. Kie.ai renders it with Veo 3 Fast
4. After two minutes the MP4 is downloaded

Who it suits

Humor channels on Shorts, Reels or TikTok that want a recurring format: a myth, a legend or a historical figure acting like a modern influencer. The prompt fixes the style. The camera is always the character’s own phone, the jokes stay deadpan and nobody wears modern clothes. A narrator, a cartoon look or a second character means rewriting the system message first.

Where to use it

Any n8n that can reach OpenAI and Kie.ai will run it. The clip comes back as a file inside the run, with the words to post it.

• n8n Cloud: import the file into a new workflow there
• Your own n8n server: the same file, the same steps
• Shorts, Reels and TikTok: download the clip and paste the agent’s title and caption
• A drive or a social account: add a node after Code to save videoMp4 or post it with the caption

Setup

Get it saves a zip with workflow.json and Mindket’s setup notes in your downloads folder. Its author puts setup at about 11 minutes.

1. In n8n, create a workflow, open its menu, choose Import from File and pick workflow.json.
2. Add an OpenAI credential and select it in OpenAI Chat Model.
3. In Generate Video2 and Get Veo3, replace <YOUR_KIE_API_KEY> after Bearer with your Kie.ai key. The brackets go too.
4. Open On form submission and switch on Required Field for Describe the video, which the form lets through empty as exported.
5. Save, send an idea through the form’s test address, then press Publish.

How to use it

Open the form from On form submission, type one line in Describe the video and send it. A character and a situation are enough. The field’s hint, Bigfoot reacting to Earth culture, shows the size. The form confirms at once, and about two minutes later the run downloads the video. Open it under Executions. Code holds the file as videoMp4, and Prompt Agent holds the title and the caption.

Example

You type a mummy trying to become a fitness coach. Prompt Agent answers in the parser’s shape, trimmed here:

{
  "title": "<up to 60 characters>",
  "post_caption": "<up to 250 characters, with emojis and hashtags>",
  "prompt": "Create a cinematic first-person selfie video … \n\nMUMMY (deadpan):\n\"…\"\n\nTime of Day: …\nLens/POV: …\nAudio (implied): …\nBackground: …\nCamera Movement: …"
}

Kie.ai renders the prompt alone. The title and the caption stay with the agent, and Code downloads the video from the first address in 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. This file puts the Kie.ai key in two headers instead, so you paste it twice. Send an idea from the form’s test address first and open each node to see what it got.

For experts. gpt-4.1 writes all three fields in one answer, and Structured Output Parser holds them to the JSON shape. Generate Video2 passes the prompt through JSON.stringify, so quotes in the dialog cannot break the request. The two Kie.ai calls use different hosts. api.kie.ai starts the render, and kieai.erweima.ai reports on it.

Code downloads with this.helpers.request. If your n8n refuses that helper, an HTTP Request node set to return a file does the same job.

Good to know

One status check decides the run.

• The check: Get Veo3 asks once, two minutes after the start. If the video is not ready, Code stops the run with the error No resultUrls[0] found on this item. Lengthen Wait Before Next Poll, or loop back to it until data.successFlag is 1
• Title and caption: no node after the agent reads them, so take them from Prompt Agent or wire them into the node that posts the clip
• The second host: if kieai.erweima.ai stops answering, give Get Veo3 the host api.kie.ai with the same path
• The original copy: the line breaks in the JSON example of the agent’s instructions had lost their backslashes and read as a bare letter n. This copy has them back
• Characters: one that a studio or a game owns can carry rights, so check before you post it
• Billing: OpenAI bills the prompt, and Kie.ai bills the render
• Your idea: it goes to OpenAI and Kie.ai, and the video sits at Kie.ai’s address before the download
• 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
• Credentials: the JSON holds none, and nothing from its author’s account