Photo and audio to talking avatar video

What it is

This n8n workflow turns one portrait and one audio recording into a talking avatar video. Fourteen nodes run in a short chain. A form takes the two files and RunningHub stores both, then a task built from them runs on a RunningHub workflow while n8n keeps asking for its status until the video is ready. The last node downloads the clip.

Who it suits

Creators and small teams who want a face that speaks their recording, without filming anyone. Skip it if you want to pick the voice, the model or the style inside n8n. The file hands the job to one fixed RunningHub workflow, so changes to the avatar itself are made on RunningHub’s side.

Where to use it

Any n8n that can reach runninghub.ai will run it. The form is the front door, so you share its address with whoever supplies the files.

• n8n Cloud: import the file into a new workflow there
• Your own n8n server: the same file and steps
• The form’s address: send it to the people who provide the portraits and recordings
• Anything after `downloadVideo`: add a node there to save the clip to a drive or mail it on

Setup

Get it saves a zip with workflow.json and Mindket’s setup notes in your downloads folder. Its author puts setup at about 5 minutes.

1. In n8n, create a workflow, open its menu, choose Import from File and pick workflow.json.
2. Open API Setup and replace YOUR_RUNNINGHUB_API with your RunningHub API key.
3. If you want, open On form submission and change the form’s title or the file types it accepts.
4. Save, open the form’s test address and send a portrait and an audio file through it.
5. Press Publish, then share the form’s production address.

How to use it

Open the form, attach a portrait and an audio file, and send them. Both fields are required. Each takes one file. The image can be .jpg, .jpeg, .png or .webp, and the audio .mp3, .wav or .flac. The form confirms at once. The run goes on behind it: upload, task, status checks, download. The video is saved in the run’s execution data, so open the run under Executions to get it.

Example

A shop owner uploads a portrait of the shop’s mascot and a recorded greeting as an .mp3. After the minute’s wait and a few status checks, downloadVideo holds a video of the portrait speaking the recording.

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. Here most boxes talk to one outside service, RunningHub, and the key you paste into API Setup is how it knows the calls are yours. Send a test through the form first. Each node then shows what it received and sent, which is where a wrong key or a rejected file type turns up.

For experts. Every call goes to one of four RunningHub endpoints, and your key travels in each request’s body.

Both uploads run side by side. The nodes Merge and Aggregate fold their replies into one item. Then createTask puts the image’s name in node 12 of the RunningHub workflow, the audio’s in node 19 and the crop in node 14. After that, Switch reads the status. SUCCESS goes to getOutput, QUEUED and RUNNING loop back through Wait1 to checkStatus, and FAILED has no line out of the switch.

Good to know

The loop has no limit, and a failed task ends the run without a word.

• Billing: RunningHub bills each task to your account
• No failure path: output 0 of Switch, for FAILED, goes nowhere, so look at the run in Executions when no video appears
• No retry limit: a task that stays queued or running is checked every 15 seconds until it finishes
• Open form: once the workflow is published, anyone with the form’s address can start a run on your RunningHub account
• Your key: it sits as plain text in API Setup, so anyone who can open the workflow can read it
• Your files: the portrait and the recording are uploaded to RunningHub, and the voice and face in them should be ones you have the right to use
• Credentials: the JSON holds none, and nothing from its author’s account