Watch the video with chapters & transcript →
AUTOMATION
In short: An n8n form collects an image plus a text description, the image is uploaded to Google Drive so the Wan API can reach it by URL, and one HTTP request kicks off the render.
Wan 2.5 generates video from a still image and a written description, and the whole flow fits in a compact n8n workflow: form in, MP4 out. Ani Björkström, a tech consultant based in Stockholm, walks through every node and gives the template away free to viewers who like the video and comment "n8n".
Two things: an image description and the image itself. The description uses a text-area field so you can write a full paragraph — this is the text sent to the API to shape the video. The image field is set to type "file", with multiple files enabled so you can submit more than one image for the video.
Because the Wan API accesses your images via URLs, not raw uploads, so the workflow first pushes the local file into Google Drive. The upload node maps the form's file field by name, pulls the file name from the schema view, and targets a chosen folder in My Drive. The connection itself needs Google Cloud credentials: at console.cloud.google.com you create a project, open APIs and services, go to the consent screen, and create a client of type web application. That produces a client ID and a client secret — save the secret immediately, because it is not shown again.
It is a single HTTP POST whose URL comes straight from the API documentation. Authentication uses a generic header credential named "authorization", with the value written as "Key", a space, then your fal API key — the key is created in fal's dashboard under API keys. The body is raw JSON carrying the prompt from the form, the Google Drive image URL, the resolution, and the duration — 10 seconds here, though 5 seconds is available and costs less.
| Node | Role in the workflow |
|---|---|
| n8n Form | Collects the image description and image file(s) |
| Google Drive upload | Stores the image so the API can fetch it by URL |
| HTTP request (POST) | Sends prompt, image URL, resolution and duration to Wan 2.5 |
| Wait + status request + if | Polls every 20 seconds until status equals completed |
| Get video (GET) | Downloads the finished MP4 from the response URL |
The workflow waits 20 seconds, calls the status endpoint, and an if-condition checks whether the status equals "completed". False loops back to the wait node for another 20 seconds — this can repeat many times, since renders take one to five minutes — and true moves on to the get-video step, which calls the response URL and returns an MP4, a format any web browser can play.
No. You can enter the Wan API URL directly in the HTTP node. fal.ai is simply convenient: one account gives access to many APIs — Veo 3, ChatGPT, Claude, Perplexity — from a single balance.
Yes. A 10-second video costs more than a 5-second one, so the duration field in the request body is also a budgeting decision.
Like the video and comment "n8n", and the workflow template is sent to you free.
Read the full video transcript, with chapters, on the video page →
Want this working inside your finance team?