Video restyler with 37 art styles

Video restyler with 37 art styles

An n8n workflow that redraws a video clip in an art style you pick from a list, such as claymation, watercolor or pixel
art. A form takes the clip, Cloudinary hosts it, and WaveSpeedAI's Wan 2.1 Ditto model renders the new version at 480p.

What is inside

• workflow.json: the workflow as n8n exports it, seven nodes and no sticky notes. It holds no credential.
• README.md: these notes.

What we changed from the original copy

• The two WaveSpeedAI calls carried Bearer <token> in their Authorization header. They now read
  Bearer <YOUR_WAVESPEED_KEY>, so the placeholder says which key belongs there.
• Nothing else. The Cloudinary address and its preset and folder were already placeholders.

Before you start

• An n8n instance, on n8n Cloud or your own server.
• A Cloudinary account with an upload preset whose signing mode is Unsigned.
• A WaveSpeedAI account with credit on it, and its API key.

Set it up

1. In n8n, create a workflow, open its menu and choose Import from File. Pick workflow.json.
2. Open Upload Video and put your cloud name in its address where it says <YOUR_CLOUD_NAME>.
3. In the same node, replace YOUR_PRESET_NAME with your unsigned preset and YOUR_FOLDER_NAME with the folder the
   source clips should land in.
4. In submitVideo and in videoStatus, replace <YOUR_WAVESPEED_KEY> in the Authorization header with your key. The
   brackets go too, so the header reads Bearer, a space and the key.
5. Save, open On form submission and send a short clip through the form's test address.
6. When the test run ends with a file in downloadVideo, press Publish.

To keep the key out of the node fields, make a Header Auth credential named Authorization with Bearer and your key as
its value. Select it under Authentication in both WaveSpeedAI nodes, then delete the typed header.

How a run goes

1. On form submission shows a form titled Video Style AI Agent with two required fields. Upload your video takes one
   file of nine video types, .mp4, .mov and .webm among them. Select the style is a dropdown of 37 styles.
2. Upload Video posts the clip to Cloudinary's video upload endpoint and gets its public url back.
3. submitVideo sends that address to WaveSpeedAI's wan-2.1/ditto endpoint. The style's label is the whole prompt.
   resolution is 480p, and seed is -1, which WaveSpeedAI reads as a random seed.
4. videoStatus asks for the job's result at once, and Switch reads its status. completed goes on to
   downloadVideo, processing goes to Wait, and failed ends the run.
5. Wait pauses 30 seconds before videoStatus asks again, as often as it takes.
6. downloadVideo fetches the first address in outputs. The new clip is a file in that node's output, under Executions.

The form only confirms that it was sent. Nothing delivers the finished clip, so add a node after downloadVideo to save it
to a drive or send it on.

Change the look

• The styles: the options of Select the style in On form submission. Each label is sent as the prompt, so keep to
  names WaveSpeedAI lists for this model.
• Sharpness: resolution in submitVideo takes 480p or 720p.
• The same result twice: give seed in submitVideo a fixed number in place of -1.
• Where the source clips go: YOUR_FOLDER_NAME in Upload Video.
• How often it checks: amount in Wait, in seconds.

Good to know

• Cost: WaveSpeedAI's page for the model, read on 2 November 2026, charges $0.20 for each 5 seconds of video at 480p and
  $0.40 at 720p. It takes clips of up to 120 seconds and bills at least 3 seconds.
• Two styles are misspelled in the original copy. IceScuIpture has a capital I where the l should be, and
  RustedlronSculpture a small L in place of the I. WaveSpeedAI lists them as IceSculpture and RustedIronSculpture,
  so correct both options before you rely on them.
• Other endings: WaveSpeedAI can also end a job as cancelled, timeout or deleted. Switch has no output for those,
  so the run stops there with an error. A failed job ends the run quietly, so look under Executions when nothing comes back.
• Open form: once the workflow is published, anyone with the form's address can spend your WaveSpeedAI credit. Turn on the
  form trigger's basic authentication or keep the address to yourself.
• Your clip: the unsigned upload makes it public on Cloudinary, and WaveSpeedAI fetches it from there.
• Style names: some labels name studios, films and toy brands (Pixar, Ghibli, Lego, Rick and Morty). Check what you may
  publish before a restyled clip goes into anything commercial.
• The key: it sits in two node headers, so take it out before you share the workflow, or move it into a credential as
  shown above.